Archaeological Informatics: Pushing the Envelope – CAA 2001: Computer Applications and Quantitative Methods in Archaeology. Proceedings of the 29th Conference, Gotland, April 2001 9781841712987, 9781407323985

Table of contents :
Cover
Title Page
Copyright
Contents
Preface
Part I
Chapter 1
Chapter 2
Chapter 3
Chapter 4
Chapter 5
Chapter 6
Chapter 7
Chapter 8
Chapter 9
Chapter 10
Chapter 11
Chapter 12
Chapter 13
Chapter 14
Chapter 15
Chapter 16
Chapter 17
Chapter 18
Chapter 19
Part II
Chapter 20
Chapter 21
Chapter 22
Chapter 23
Chapter 24
Chapter 25
Part III
Chapter 26
Chapter 27
Chapter 28
Chapter 29
Part IV
Chapter 30
Chapter 31
Chapter 32
Chapter 33
Chapter 34
Chapter 35
Chapter 36
Chapter 37
Chapter 38
Chapter 39
Part V
Chapter 40
Chapter 41
Chapter 42
Chapter 43
Chapter 44
Chapter 45
Chapter 46
Chapter 47
Chapter 48
Chapter 49
Chapter 50
Chapter 51
Chapter 52
Part VI
Chapter 53
Chapter 54
Chapter 55
Chapter 56
Chapter 57
Chapter 58
Chapter 59
Chapter 60
Chapter 61
Chapter 62
Chapter 63
Chapter 64
Chapter 65
Chapter 66
Chapter 67
Part VII
Chapter 68
Chapter 69
Chapter 70
Chapter 71
Chapter 72

Citation preview

l na tio ne di nli ad l o ith ria W ate m

BAR  S1016  2002   BURENHULT (Ed.)   ARCHAEOLOGICAL INFORMATICS: PUSHING THE ENVELOPE CAA 2001

Archaeological Informatics: Pushing the Envelope CAA 2001 Computer Applications and Quantitative Methods in Archaeology Proceedings of the 29th Conference, Gotland, April 2001

Edited by

Göran Burenhult co-editor

Johan Arvidsson

BAR International Series 1016 9 781841 712987

B A R

2002

Archaeological Informatics: Pushing the Envelope CAA 2001

Archaeological Informatics: Pushing the Envelope CAA 2001

Computer Applications and Quantitative Methods in Archaeology

Computer Applications and Quantitative Methods in Archaeology

Proceedings of the 29th Conference, Gotland, April 2001

Proceedings of the 29th Conference, Gotland, April 2001

Edited by

Edited by

Göran Burenhult

Göran Burenhult

co-editor

co-editor

Johan Arvidsson

Johan Arvidsson

BAR International Series 1016 2002

BAR International Series 1016 2002

Published in 2016 by BAR Publishing, Oxford BAR International Series 1016 Archaeological Informatics: Pushing the Envelope - CAA. 2001

© The editors and contributors severally and the Publisher 2002 The authors' moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781841712987 paperback ISBN 9781407323985 e-format DOI https://doi.org/10.30861/9781841712987 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd/ Hadrian Books Ltd, the Series principal publisher, in 2002. This present volume is published by BAR Publishing, 2016.

BAR

PUBLISHING BAR titles are available from:

EMAIL

PHONE FAX

BAR Publishing 122 Banbury Rd, Oxford, OX2 ?BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Contents Preface ................................................................................................................................................................................................................ 1

Papers in the GIS category

Contents Preface ................................................................................................................................................................................................................ 1

Papers in the GIS category

Extracting “Natural Pathways” from a Digital Elevation Model.................................................................................................................. 5 Applications to Landscape Archaeological Studies Gino Bellavia

Extracting “Natural Pathways” from a Digital Elevation Model.................................................................................................................. 5 Applications to Landscape Archaeological Studies Gino Bellavia

2000 Years of Town Planning in Vienna ....................................................................................................................................................... 13 Wolfgang Börner

2000 Years of Town Planning in Vienna ....................................................................................................................................................... 13 Wolfgang Börner

A 5000-Years History of Settlement and Irrigation in the Murghab Delta (Turkmenistan) ....................................................................................................................................................... 21 An Attempt of Reconstruction of Ancient Deltaic System Barbara Cerasetti

A 5000-Years History of Settlement and Irrigation in the Murghab Delta (Turkmenistan) ....................................................................................................................................................... 21 An Attempt of Reconstruction of Ancient Deltaic System Barbara Cerasetti

Building, building on the wall…..................................................................................................................................................................... 29 A reflection of actual building dimensions? Dora Constantinidis

Building, building on the wall…..................................................................................................................................................................... 29 A reflection of actual building dimensions? Dora Constantinidis

Historic landscape assessment: ...................................................................................................................................................................... 35 The East of England Experience Lynn Dyson-Bruce

Historic landscape assessment: ...................................................................................................................................................................... 35 The East of England Experience Lynn Dyson-Bruce

GIS approach to Iberian iron age landscape in central-south Valencia region (Spain) ..................................................................................................................................................... 43 Ignacio Grau Mira

GIS approach to Iberian iron age landscape in central-south Valencia region (Spain) ..................................................................................................................................................... 43 Ignacio Grau Mira

Developing an Information System for Archaeological Sites and Monuments - Data Model and Construction..................................... 49 David Haskiya

Developing an Information System for Archaeological Sites and Monuments - Data Model and Construction..................................... 49 David Haskiya

Using viewsheds wisely: developing sound methodologies from spatial analyses of megalithic monuments in western Scotland......... 53 Gail Higginbottom, Ken Simpson and Roger Clay

Using viewsheds wisely: developing sound methodologies from spatial analyses of megalithic monuments in western Scotland......... 53 Gail Higginbottom, Ken Simpson and Roger Clay

Reconstructing the seascape at the mouth of the Oder ............................................................................................................................... 63 Elaboration of a DBM-model based on 1912-soundings George Indruszewski

Reconstructing the seascape at the mouth of the Oder ............................................................................................................................... 63 Elaboration of a DBM-model based on 1912-soundings George Indruszewski

The TimeMap Kiosk: Delivering Historical Images in a Spatio-Temporal Context ......................................................................................................................................................................... 71 Ian Johnson and Andrew Wilson

The TimeMap Kiosk: Delivering Historical Images in a Spatio-Temporal Context ......................................................................................................................................................................... 71 Ian Johnson and Andrew Wilson

The answer is blowin’ in the wind. Research desires and data possibilities. ............................................................................................................................................................................... 79 Hans Kamermans

The answer is blowin’ in the wind. Research desires and data possibilities. ............................................................................................................................................................................... 79 Hans Kamermans

A hundred years of lake contour fluctuation in the Hamun-i Helmand: A GIS based system for the study and the recovery of archaeological information in the Iranian Sistan (1899-1999)............................................................................................................... 85 Sabatino Laurenza and Sophie Pornet-Laurenza

A hundred years of lake contour fluctuation in the Hamun-i Helmand: A GIS based system for the study and the recovery of archaeological information in the Iranian Sistan (1899-1999)............................................................................................................... 85 Sabatino Laurenza and Sophie Pornet-Laurenza

From Stratigraphic Unit to the mouse: a GIS based system for the excavation of historical complex. The case study of Pompeii....... 93 Sabatino Laurenza and Cristiano Putzolu

From Stratigraphic Unit to the mouse: a GIS based system for the excavation of historical complex. The case study of Pompeii....... 93 Sabatino Laurenza and Cristiano Putzolu

Statistical analysis of the distribution of modern primates: a comparative approach to the spatial analysis of the Palaeolithic ........ 105 Katharine MacDonald

Statistical analysis of the distribution of modern primates: a comparative approach to the spatial analysis of the Palaeolithic ........ 105 Katharine MacDonald

The Creation and Potential Applications of a 3-Dimensional GIS for the Early Hominid Site of Swartkrans, South Africa ............. 113 Joseph D. Nigro, W. Fredrick Limp, Kenneth K. Kvamme, Darryl J. de Ruiter and Lee R. Berger

The Creation and Potential Applications of a 3-Dimensional GIS for the Early Hominid Site of Swartkrans, South Africa ............. 113 Joseph D. Nigro, W. Fredrick Limp, Kenneth K. Kvamme, Darryl J. de Ruiter and Lee R. Berger

GIS contribution to urban history and to the reconstruction of ancient landscape................................................................................. 125 Sofia Pescarin

GIS contribution to urban history and to the reconstruction of ancient landscape................................................................................. 125 Sofia Pescarin

i

i

GIS Modeling of the Minoan Peak Sanctuaries of East Crete ................................................................................................................... 129 S. Soetens, A. Sarris, S. Topouzi and A. Tripolitsiotis

GIS Modeling of the Minoan Peak Sanctuaries of East Crete ................................................................................................................... 129 S. Soetens, A. Sarris, S. Topouzi and A. Tripolitsiotis

Modeling Archaeological and Historical Cognitive Landscapes in the Greater Yellowstone Region (Wyoming, Montana, and Idaho, USA) Using Geographic Information Systems................................................................. 139 Thomas G. Whitley

Modeling Archaeological and Historical Cognitive Landscapes in the Greater Yellowstone Region (Wyoming, Montana, and Idaho, USA) Using Geographic Information Systems................................................................. 139 Thomas G. Whitley

GIS and Space Analysis in the study of the Hospitallers’ fortifications in the Dodecanese .................................................................... 149 Nicholas Zarifis and Despina Brokou

GIS and Space Analysis in the study of the Hospitallers’ fortifications in the Dodecanese .................................................................... 149 Nicholas Zarifis and Despina Brokou

Papers in the Virtual Archaeology category

Papers in the Virtual Archaeology category

Reconstructions of the Excavations of Two Iron Age Chariot Burials from Belgium ............................................................................. 157 Applying Virtual Reality to Old Excavation Data Geoff Avern

Reconstructions of the Excavations of Two Iron Age Chariot Burials from Belgium ............................................................................. 157 Applying Virtual Reality to Old Excavation Data Geoff Avern

3D standards for scientific communication ................................................................................................................................................. 163 Francesca Cantone

3D standards for scientific communication ................................................................................................................................................. 163 Francesca Cantone

Acoustic reconstruction of music performance spaces using three-dimentional digital waveguide mesh models................................. 173 Guilherme Campos, David M. Howard and Steve Dobson

Acoustic reconstruction of music performance spaces using three-dimentional digital waveguide mesh models................................. 173 Guilherme Campos, David M. Howard and Steve Dobson

A virtual journey through a Roman settlement. Aloria ............................................................................................................................. 177 Juan José Fuldain González

A virtual journey through a Roman settlement. Aloria ............................................................................................................................. 177 Juan José Fuldain González

Topometrical measurements in Tiryns, Greece .......................................................................................................................................... 181 Report on a co-operate project between physics and archaeology Maria Shinoto, Zoltán Böröcz, Carsten Thomas, Dieter Dirksen, Joseph Maran and Gert von Bally

Topometrical measurements in Tiryns, Greece .......................................................................................................................................... 181 Report on a co-operate project between physics and archaeology Maria Shinoto, Zoltán Böröcz, Carsten Thomas, Dieter Dirksen, Joseph Maran and Gert von Bally

Advances in Geometric Modeling and Feature Extraction on Pots, Rocks and Bones for Representation and Query via the Internet..................................................................................................................................................................................................... 191 Utsav A. Schurmans, Anshuman Razdan, Arleyn Simon, Mary Marzke, Peter McCartney, David Van Alfen, Gram Jones, Mary Zhu, Dezhi Liu, Myungsoo Bae, Jeremy Rowe Gerald Farin and Dan Collins

Advances in Geometric Modeling and Feature Extraction on Pots, Rocks and Bones for Representation and Query via the Internet..................................................................................................................................................................................................... 191 Utsav A. Schurmans, Anshuman Razdan, Arleyn Simon, Mary Marzke, Peter McCartney, David Van Alfen, Gram Jones, Mary Zhu, Dezhi Liu, Myungsoo Bae, Jeremy Rowe Gerald Farin and Dan Collins

Papers in the Osteology category

Papers in the Osteology category

Neural Network Classification of Skeletal Remains ................................................................................................................................... 205 Suzanne Bell and Richard Jantz

Neural Network Classification of Skeletal Remains ................................................................................................................................... 205 Suzanne Bell and Richard Jantz

Osteology for the new millennium................................................................................................................................................................ 213 Gustav Jonsson

Osteology for the new millennium................................................................................................................................................................ 213 Gustav Jonsson

3-D CAT-scan: Anthropology, Archaeology and Virtual Reality.............................................................................................................. 217 Niels Lynnerup

3-D CAT-scan: Anthropology, Archaeology and Virtual Reality.............................................................................................................. 217 Niels Lynnerup

The Statistics of Archaeological Deformation Processes ........................................................................................................................... 221 An archaeozoological experiment Laura Mameli, Juan A. Barcelo and Jordi Estevez

The Statistics of Archaeological Deformation Processes ........................................................................................................................... 221 An archaeozoological experiment Laura Mameli, Juan A. Barcelo and Jordi Estevez

Papers in the Internet Applications and Cultural Heritage Management categories

Papers in the Internet Applications and Cultural Heritage Management categories

Generating Multilingual Personalized Descriptions of Museum Exhibits – The M-PIRO Project ...................................................................................................................................................................................... 233 Ion Androutsopoulos, Aggeliki Dimitromanolaki, Vassiliki Kokkinaki, Jo Calder, Jon Oberlander and Elena Not

Generating Multilingual Personalized Descriptions of Museum Exhibits – The M-PIRO Project ...................................................................................................................................................................................... 233 Ion Androutsopoulos, Aggeliki Dimitromanolaki, Vassiliki Kokkinaki, Jo Calder, Jon Oberlander and Elena Not

ii

ii

Joined up writing: an Internet portal for research into the Historic Environment ................................................................................. 243 Tony Austin, Francisco Pinto, Julian Richards and Nick Ryan

Joined up writing: an Internet portal for research into the Historic Environment ................................................................................. 243 Tony Austin, Francisco Pinto, Julian Richards and Nick Ryan

The Archaeological Map of Egypt................................................................................................................................................................ 253 (Archaeological Heritage Resource Management System) Fathi Saleh and Reem Bahgat

The Archaeological Map of Egypt................................................................................................................................................................ 253 (Archaeological Heritage Resource Management System) Fathi Saleh and Reem Bahgat

Geographic Information Systems and Archaeology: the case of ancient Nora (Pula-Cagliari)................................................................................................................................................ 261 Anna Maria Colavitti and Giancarlo Deplano

Geographic Information Systems and Archaeology: the case of ancient Nora (Pula-Cagliari)................................................................................................................................................ 261 Anna Maria Colavitti and Giancarlo Deplano

XML Encoding of Archaeological Unstructured Data ............................................................................................................................... 267 Marco Crescioli, Andrea D’Andrea and Franco Niccolucci

XML Encoding of Archaeological Unstructured Data ............................................................................................................................... 267 Marco Crescioli, Andrea D’Andrea and Franco Niccolucci

Cultural landscape, computers and characterisation: ............................................................................................................................... 277 GIS-based Historic Landscape Characterisation as a tool for archaeological resource management in England Graham Fairclough

Cultural landscape, computers and characterisation: ............................................................................................................................... 277 GIS-based Historic Landscape Characterisation as a tool for archaeological resource management in England Graham Fairclough

Corinth Computer Project: Internet Education ......................................................................................................................................... 295 David Gilman Romano and Nicholas L. Stapp

Corinth Computer Project: Internet Education ......................................................................................................................................... 295 David Gilman Romano and Nicholas L. Stapp

Desktop - Photogrammetry and its Link to Web Publishing ..................................................................................................................... 301 Günter Pomaska

Desktop - Photogrammetry and its Link to Web Publishing ..................................................................................................................... 301 Günter Pomaska

A Web-based Digital Archaeological Map of Lasithi, E. Crete. ................................................................................................................ 309 Apostolos Sarris, Katerina Bichta, Marina Giasta, Anthi Giourou, Evaggelia Karimali, Vaggelis Kevgas, Kostas Margetousakis, Eleni Peraki, Steven Soetens, Katerina Tzaneteas, SofiaTopouzi and Achilleas Tripolitsiotis

A Web-based Digital Archaeological Map of Lasithi, E. Crete. ................................................................................................................ 309 Apostolos Sarris, Katerina Bichta, Marina Giasta, Anthi Giourou, Evaggelia Karimali, Vaggelis Kevgas, Kostas Margetousakis, Eleni Peraki, Steven Soetens, Katerina Tzaneteas, SofiaTopouzi and Achilleas Tripolitsiotis

The Archaeologist Files: An approach to the digital contextualization of archaeological finds in user adaptive information systems ..................................................................................................................................................................................................... 325 Isto Vatanen

The Archaeologist Files: An approach to the digital contextualization of archaeological finds in user adaptive information systems ..................................................................................................................................................................................................... 325 Isto Vatanen

Papers in the Survey and Mapping, Archaeometry, GPS and CAD categories

Papers in the Survey and Mapping, Archaeometry, GPS and CAD categories

New Tools for Understanding Plains Indian Sites in Grasslands National Park, Canada ...................................................................... 335 Gary Adams

New Tools for Understanding Plains Indian Sites in Grasslands National Park, Canada ...................................................................... 335 Gary Adams

Archaeological Survey as Optimal Search .................................................................................................................................................. 341 E. B. Banning

Archaeological Survey as Optimal Search .................................................................................................................................................. 341 E. B. Banning

Archaeological spatial modelling.................................................................................................................................................................. 351 A case study from Beagle Channel (Argentina) Juan A. Barcelo, Ernesto L. Piana and Daniel R. Martinioni

Archaeological spatial modelling.................................................................................................................................................................. 351 A case study from Beagle Channel (Argentina) Juan A. Barcelo, Ernesto L. Piana and Daniel R. Martinioni

Views on „castrum cuscin“ ........................................................................................................................................................................... 361 Archeology and archaeometry under water Ralf Bleile, Cornelius Meyer and Burkhardt Ullrich

Views on „castrum cuscin“ ........................................................................................................................................................................... 361 Archeology and archaeometry under water Ralf Bleile, Cornelius Meyer and Burkhardt Ullrich

A PC-based stereoscopic measurement system for the generation of digital object models.................................................................... 371 Frank Boochs, Stephan Eckhardt and Ben Fischer

A PC-based stereoscopic measurement system for the generation of digital object models.................................................................... 371 Frank Boochs, Stephan Eckhardt and Ben Fischer

Infra-red (Reflectance) Spectroscopy of Ceramics from Tell Halawa, Syria ........................................................................................... 379 Murray Lee Eiland and Quentin Williams

Infra-red (Reflectance) Spectroscopy of Ceramics from Tell Halawa, Syria ........................................................................................... 379 Murray Lee Eiland and Quentin Williams

Detailed Topography and Surface Survey. What is the point? Tanagra City Survey 2000 .................................................................... 385 Emeri Farinetti and Lefteris Sigalos

Detailed Topography and Surface Survey. What is the point? Tanagra City Survey 2000 .................................................................... 385 Emeri Farinetti and Lefteris Sigalos

The application of high-resolution satellite imagery for the detection of ancient Minoan features on Crete ....................................................................................................................................................................... 393 Lefki Pavlidis, Clive S Fraser and Cliff Ogleby

The application of high-resolution satellite imagery for the detection of ancient Minoan features on Crete ....................................................................................................................................................................... 393 Lefki Pavlidis, Clive S Fraser and Cliff Ogleby

iii

iii

Educating the Digital Fieldwork Assistant .................................................................................................................................................. 401 Nick Ryan and Martijn van Leusen

Educating the Digital Fieldwork Assistant .................................................................................................................................................. 401 Nick Ryan and Martijn van Leusen

From Field Books to Powerbook .................................................................................................................................................................. 413 Computer Applications and the Promontory Palace of Herod the Great James G. Schryver

From Field Books to Powerbook .................................................................................................................................................................. 413 Computer Applications and the Promontory Palace of Herod the Great James G. Schryver

Errors & Inaccuracies in Repositioning of Archaeological Sites............................................................................................................... 417 S. Topouzi, A. Tripolitsiotis, A. Sarris, S. Mertikas and S. Soetens

Errors & Inaccuracies in Repositioning of Archaeological Sites............................................................................................................... 417 S. Topouzi, A. Tripolitsiotis, A. Sarris, S. Mertikas and S. Soetens

Three-dimensional excavation plans and 3D Studio Max .......................................................................................................................... 427 Experiences from the excavations of the medieval town of Naantali, Finland Kari Uotila and Carita Tulkki

Three-dimensional excavation plans and 3D Studio Max .......................................................................................................................... 427 Experiences from the excavations of the medieval town of Naantali, Finland Kari Uotila and Carita Tulkki

Handling digital 3-D record of archaeological excavation data ................................................................................................................ 431 Mikhail Zhukovsky

Handling digital 3-D record of archaeological excavation data ................................................................................................................ 431 Mikhail Zhukovsky

Papers in the Database Applications and Statistics and Quantitative Methods categories

Papers in the Database Applications and Statistics and Quantitative Methods categories

Approaches to petrographic data analysis using S-Plus............................................................................................................................. 441 C.C. Beardah, M.J. Baxter, I. Papageorgiou and M.A. Cau

Approaches to petrographic data analysis using S-Plus............................................................................................................................. 441 C.C. Beardah, M.J. Baxter, I. Papageorgiou and M.A. Cau

Developing an information system for archaeological sites and monuments – administration and maintenance model .......................................................................................................................................................................... 449 Malin Blomqvist and Cissela Génetay-Lindholm

Developing an information system for archaeological sites and monuments – administration and maintenance model .......................................................................................................................................................................... 449 Malin Blomqvist and Cissela Génetay-Lindholm

How can a database full of Bugs help reconstruct the climate?................................................................................................................. 453 Phil Buckland and Paul Buckland

How can a database full of Bugs help reconstruct the climate?................................................................................................................. 453 Phil Buckland and Paul Buckland

Pattern recognition applied to Rock Art ..................................................................................................................................................... 463 Diego Diaz and Damián Castro

Pattern recognition applied to Rock Art ..................................................................................................................................................... 463 Diego Diaz and Damián Castro

Creating of a database for prehistoric sites: which are the goals, the strategy and what means to put in place? ............................................................................................................................................................ 469 Pierre Corboud

Creating of a database for prehistoric sites: which are the goals, the strategy and what means to put in place? ............................................................................................................................................................ 469 Pierre Corboud

The Computer Catalogue of the Kunstkammer Museum Collections and perspectives of an Internet-shared Anthropological Database...................................................................................................................................... 475 Yuri Osintsov, Yuri Chistov and Dmitriy Gerasimov

The Computer Catalogue of the Kunstkammer Museum Collections and perspectives of an Internet-shared Anthropological Database...................................................................................................................................... 475 Yuri Osintsov, Yuri Chistov and Dmitriy Gerasimov

Reclaiming Old Data: The Wasden Site Research Project ........................................................................................................................ 483 E. S. Lohse and J. Anderson

Reclaiming Old Data: The Wasden Site Research Project ........................................................................................................................ 483 E. S. Lohse and J. Anderson

The Archaeology-Palaeobotany-Palynology Database on the Palaeolithic, Mesolithic and early Neolithic sites of the Former USSR Area .............................................................................................................................................................................................. 491 Yuri Stepanov, Galina Levkovskaya, Mikhail Anikovitch, Nikolai Anisutkin, Elena Beliaeva, Vladimir Shumkin, Veronika Stegantzeva, Vladimir Timofeev, Galina Sinitsyna, Anastassia Bogolubova and Andrey Stegantzev

The Archaeology-Palaeobotany-Palynology Database on the Palaeolithic, Mesolithic and early Neolithic sites of the Former USSR Area .............................................................................................................................................................................................. 491 Yuri Stepanov, Galina Levkovskaya, Mikhail Anikovitch, Nikolai Anisutkin, Elena Beliaeva, Vladimir Shumkin, Veronika Stegantzeva, Vladimir Timofeev, Galina Sinitsyna, Anastassia Bogolubova and Andrey Stegantzev

Never under-estimate the power of a model................................................................................................................................................ 495 Clive Robert Orton

Never under-estimate the power of a model................................................................................................................................................ 495 Clive Robert Orton

Deriving ancient foot units from building dimensions: a statistical approach employing cosine quantogram analysis ................................................................................................................................................. 501 Jari Pakkanen

Deriving ancient foot units from building dimensions: a statistical approach employing cosine quantogram analysis ................................................................................................................................................. 501 Jari Pakkanen

A Genetic Algorithm problem solver for Archaeology............................................................................................................................... 507 Carlos Reynoso and Edward Jezierski

A Genetic Algorithm problem solver for Archaeology............................................................................................................................... 507 Carlos Reynoso and Edward Jezierski

Visualising the Neolithic transition in Europe ............................................................................................................................................ 511 Thembi Russell and James Steele

Visualising the Neolithic transition in Europe ............................................................................................................................................ 511 Thembi Russell and James Steele

iv

iv

Archaeological databases: what are they and what do they mean? .......................................................................................................... 517 Robert Schlader

Archaeological databases: what are they and what do they mean? .......................................................................................................... 517 Robert Schlader

The Use of Permutations to Explain the Hackness Cross Tree Rune Inscription .................................................................................... 521 Richard Sermon

The Use of Permutations to Explain the Hackness Cross Tree Rune Inscription .................................................................................... 521 Richard Sermon

The computer as a tool in a dialectical approach to data processing ........................................................................................................ 527 Maria Shinoto

The computer as a tool in a dialectical approach to data processing ........................................................................................................ 527 Maria Shinoto

Papers in the Posters and Workshops categories

Papers in the Posters and Workshops categories

Presentation of a Database for funerary Analysis and Proposals for its Increments and Developments ............................................................................................................................................................... 543 Francesca Fulminante

Presentation of a Database for funerary Analysis and Proposals for its Increments and Developments ............................................................................................................................................................... 543 Francesca Fulminante

A GIS Application for the study on water supply and draining system in the ancient capital cities in Japan ...................................................................................................................................................... 547 Akihiro Kaneda and Susumu Morimoto

A GIS Application for the study on water supply and draining system in the ancient capital cities in Japan ...................................................................................................................................................... 547 Akihiro Kaneda and Susumu Morimoto

Bjäre and Bowland: Computer Applications in European Pathways to Cultural Landscapes, a Culture 2000 programme .................................................................................................................................................................... 551 Jenny Nord Paulsson and Graham Fairclough

Bjäre and Bowland: Computer Applications in European Pathways to Cultural Landscapes, a Culture 2000 programme .................................................................................................................................................................... 551 Jenny Nord Paulsson and Graham Fairclough

Ancient Mantineia’s Defence Network Reconsidered Through a GIS Approach ................................................................................... 559 S. Topouzi, A. Sarris, Y. Pikoulas, S. Mertikas, X. Frantzis and A. Giourou

Ancient Mantineia’s Defence Network Reconsidered Through a GIS Approach ................................................................................... 559 S. Topouzi, A. Sarris, Y. Pikoulas, S. Mertikas, X. Frantzis and A. Giourou

Laser Scanning and Digital Close Range Photogrammetry for Capturing 3-d Archaeological Objects: a Comparison of Quality and Practicality ................................................................................................... 567 Athanasios Velios and John P. Harrison

Laser Scanning and Digital Close Range Photogrammetry for Capturing 3-d Archaeological Objects: a Comparison of Quality and Practicality ................................................................................................... 567 Athanasios Velios and John P. Harrison

CAA 2001 Abstracts on CD

CAA 2001 Abstracts on CD

Abstracts in the GIS category .....................................................................................................................................................1

Abstracts in the GIS category .....................................................................................................................................................1

Abstracts in the Virtual Reality and Digital Image Processing categories............................................................................17

Abstracts in the Virtual Reality and Digital Image Processing categories............................................................................17

Abstracts in the Osteology category..........................................................................................................................................26

Abstracts in the Osteology category..........................................................................................................................................26

Abstracts in the Internet Applications and Cultural Heritage Management categories......................................................30

Abstracts in the Internet Applications and Cultural Heritage Management categories......................................................30

Abstracts in the Survey and Mapping, Archaeometry, GPS and CAD categories ...............................................................43

Abstracts in the Survey and Mapping, Archaeometry, GPS and CAD categories ...............................................................43

Abstracts in the Database Applications and Statistics and Quantitative Methods categories.............................................51

Abstracts in the Database Applications and Statistics and Quantitative Methods categories.............................................51

Abstracts to workshops..............................................................................................................................................................61

Abstracts to workshops..............................................................................................................................................................61

Abstracts to posters ....................................................................................................................................................................69

Abstracts to posters ....................................................................................................................................................................69

Please note that the CD referred to above has now been replaced with a download available at www.barpublishing.com/additional-downloads.html

v

v

vi

vi

Preface

Preface

This volume presents the proceedings of the 29th CAA conference, entitled Archaeological Informatics – Pushing the Envelope. Computer Applications and Quantitative Methods in Archaeology. The conference was held in Visby, on the island of Gotland, in Sweden, April 25-29, 2001. Altogether, 289 delegates from 30 countries participated in the conference. A total of 119 papers were presented, together with 19 posters and 5 workshops. In this volume, 72 papers are presented. All illustrations are printed in the volume in b/w, also those submitted as colour images, as well as on the enclosed CD-ROM disc, on which the colour images can be studied in true colours. Abstracts to all presented papers, as well as posters and workshops, are also included.

This volume presents the proceedings of the 29th CAA conference, entitled Archaeological Informatics – Pushing the Envelope. Computer Applications and Quantitative Methods in Archaeology. The conference was held in Visby, on the island of Gotland, in Sweden, April 25-29, 2001. Altogether, 289 delegates from 30 countries participated in the conference. A total of 119 papers were presented, together with 19 posters and 5 workshops. In this volume, 72 papers are presented. All illustrations are printed in the volume in b/w, also those submitted as colour images, as well as on the enclosed CD-ROM disc, on which the colour images can be studied in true colours. Abstracts to all presented papers, as well as posters and workshops, are also included.

The conference was officially opened by Princess Christina, Mrs. Magnuson, in the Picture Stone Hall at the Gotland Historical Museum. The conference was introduced by an invited lecture, presented by Associate Professor Paul Sinclair and Dr. Markku Pyykönen, Computer aided assessments of human responses and contributions to environmental change in Africa over the last 2600 years, after which the conference was split up into parallel sessions and workshops. Titles of the sessions were:

The conference was officially opened by Princess Christina, Mrs. Magnuson, in the Picture Stone Hall at the Gotland Historical Museum. The conference was introduced by an invited lecture, presented by Associate Professor Paul Sinclair and Dr. Markku Pyykönen, Computer aided assessments of human responses and contributions to environmental change in Africa over the last 2600 years, after which the conference was split up into parallel sessions and workshops. Titles of the sessions were:

-

-

GIS Virtual Archaeology Osteology Internet Applications and Cultural Heritage Management Survey and Mapping, Archaeometry, GPS and CAD Database Applications and Statistical and Quantitative Methods

GIS Virtual Archaeology Osteology Internet Applications and Cultural Heritage Management Survey and Mapping, Archaeometry, GPS and CAD Database Applications and Statistical and Quantitative Methods

All sessions were chaired by internationally recognized experts, and those experts also acted as peer reviewers. Despite these reviewers being anonymous in the individual case, we would like to acknowledge them for their work, and also thank them for prompt and clear comments and advice to the various authors: Juan Barcelo, Martin Doerr, Ebba During, Emeri Farinetti, Ian Johnson, Hans Kamermans, Franco Niccolucci, Clive Orton, David Gilman Romano, Nick Ryan, Apostolos Sarris and Steve Stead.

All sessions were chaired by internationally recognized experts, and those experts also acted as peer reviewers. Despite these reviewers being anonymous in the individual case, we would like to acknowledge them for their work, and also thank them for prompt and clear comments and advice to the various authors: Juan Barcelo, Martin Doerr, Ebba During, Emeri Farinetti, Ian Johnson, Hans Kamermans, Franco Niccolucci, Clive Orton, David Gilman Romano, Nick Ryan, Apostolos Sarris and Steve Stead.

We would also like to thank everyone who made the organization in Visby possible during the five hectic days in April, including individual students and the students’ organization RINDI, the International Secretary, the Gotland University College technical and administrative staff, and the Historical Museum of Gotland. A special acknowledgement goes to Gustaf Svedjemo, for keeping total and successful control of the computer equipment during the conference. For the English language correction of the papers to be published, we would like to thank Mrs. Cathy Gow Sjöblom.

We would also like to thank everyone who made the organization in Visby possible during the five hectic days in April, including individual students and the students’ organization RINDI, the International Secretary, the Gotland University College technical and administrative staff, and the Historical Museum of Gotland. A special acknowledgement goes to Gustaf Svedjemo, for keeping total and successful control of the computer equipment during the conference. For the English language correction of the papers to be published, we would like to thank Mrs. Cathy Gow Sjöblom.

For generous funding of the conference, we would like to thank the European Union – the Structural Funds (Region Objective 2 Islands Sweden), the Gotland Regional Development Unit, Gotland County Council, the Municipality of Gotland, Sällskapet DBV and Trimble AB.

For generous funding of the conference, we would like to thank the European Union – the Structural Funds (Region Objective 2 Islands Sweden), the Gotland Regional Development Unit, Gotland County Council, the Municipality of Gotland, Sällskapet DBV and Trimble AB.

Göran Burenhult Johan Arvidsson

Göran Burenhult Johan Arvidsson

Visby, January 2002

Visby, January 2002

1

1

Papers in the GIS category

Extracting "Natural Pathways" from a Digital Elevation Model Applications to Landscape Archaeological Studies. Gino Bellavia Ancient History & Archaeology University of Birmingham Edgbaston Birmingham, B 15 2TT Phone: +44 24 764 16994 - E-mail: [email protected]

Abstract: The relationship between the distribution of archaeological sites and "natural pathways" through the landscape forms the first stage of a research programme into the relationships between prehistoric "ritual" sites in the UK and their landscape environment. A technique has been devised to extract "natural" pathways, from a landscape Digital Elevation Model. The method is derived from GIS based hydrological modelling techniques and identifies potential multi-scale pathways across a landscape. The "natural pathways" modelling process has been used on several areas in the UK to test the validity of the approach. Initial tests have produced promising results and indicates the usefulness and validity of the methodology. Further analysis on an area around the prehistoric site of Stonehenge hint at several possible relationships between burial monuments and "natural pathways" in the landscape. Key words: Natural Pathways, Digital Elevation Models, Cost Surfaces, Prehistoric routes, Ritual sites.

Introduction

pathways also provide a human structure to the landscape and allow a direct connection to the landscape and its elements. Their existence and nature should be an important element in archaeological interpretation. As Ingold ( 1993) states, "In short, the landscape is the world as it is known to those who dwell therein, who inhabit its places and journey along the paths connecting them".

Overall, archaeological landscape studies have focussed on the potential relationships between sites and their landscape environment on a site by site basis, or by small-scale regional studies. While these studies may enhance the interpretation of these sites and perhaps regional patterns, they do not provide a higher level view of site characteristics. For our research, newly available data sources were to be used to facilitate a landscape analysis at the macro level. Specifically, the investigation aims to look at generalised patterns for the whole of the UK.

Potential relationships between prehistoric sites and pathways have been mentioned by many authors, however the locations of these pathways are often subjectively defined. In (Llobera, 1996; ParceroOubina et al. 1998) the authors suggest that rock art and megalithic barrow sites in Iberia, are related to probable pathways. (BRADLEY et al. 1994) have looked at the relationship between potential pathways and rock art sites, however the notion of "paths" is again, subjective. Bodmin Moor and its cairns, are suggested by (Tilley, 1996), to be accessed by pathways along the main river courses and their tributaries.

Initial researches into the patterns of site location focussed on relatively simple, 2-dimensional distribution analyses. These early studies generally treated the landscape as a flat plane. Recently, Digital Elevation Models (DEM) have facilitated 3dimensional analyses. The most commonly analytically studied aspects of the landscape and sites have been intervisibility and viewshed type analyses. These techniques aim to provide an insight into site distribution and form, but may fail to take account of other, formalised aspects of the landscape, such as pathways.

(Loveday, 1998) calculated a correlation, greater than 54%, between the alignments of Roman roads and nearby double entrance henges. The author comments on the possible reasons for this correlation, which in some cases may relate to the natural topography, indicating probable pathways through the henge openings. In this scenario, the Roman roads simply follow the same terrain, as might a modem pathway. Although the alignments also occur in relatively flat terrain, which is difficult to interpret, the data set is sufficient to hint at a potential relationship between the alignments of the road and double entrance henges.

It has been stated that settlements are often located at natural crossroads or thoroughfares, suggesting that before intensive road building schemes took place, sites were situated close to natural pathways through the landscape.

The formation of tracks and pathways, at any period, on a landscape, is an inevitable result of human and animal movements over time. These paths may permeate the landscape for long periods, waxing and waning in their use and significance. These

5

The locations of road networks are generally understood during the Roman period. Whilst a recent study 1 on the Ridgeway path in England postulates a prehistoric date for that pathway, many studies of prehistoric pathways suffer from a dearth of independent evidence. Furthermore, their dating is often concluded from the prevalence of nearby prehistoric sites.

modelling. The essential difference is the nature of the input data. Whilst hydrological analysis uses a raw DEM to produce a stream model, the extraction of "natural pathways" uses a constraint surface model as input. All of the analytical stages are cell based, as the terrain models used in this research are grid-based.

In order to ascertain any correlation between sites and "natural pathways", a formal methodology to extract potential "natural pathways" independently from site location had to be developed.

Further information on the separate stages can be found in Arcview 3.2 GIS (spatial analyst & hydrological analysis vl extensions). These discuss a notion of flow as they relate to hydrology. This nomenclature is used here but it is useful to think of flow, in relation to "natural pathways", as equivalent to potential movements of people.

At the scale of the analysis, discussed later, the natural topography of the UK is assumed not to have significantly changed since the prehistoric period. Therefore a modem landscape Digital Elevation Model forms the basic data set for deriving these "natural pathways".

The Generation of a Constraint or Cost Surface Grid The input to the path network algorithm is a cost grid defining the relative frictions of human traversal through each grid cell that depends on the type of landscape component in that cell.

The "natural pathways" are defined by the natural constraints of traversal across a landscape, and in general, will be routes of minimum constraint. It is important to note that the concept of natural pathways does not explicitly imply their human use as a path, but only their potential to be used as such.

The desired constraint surface may comprise a number oflandscape and/or perceptive elements such as; Rivers

Review of Path Finding Techniques

Navigable rivers Rocky mountain streams

Almost all studies on the extraction of paths from a landscape relate to finding optimal least cost paths between known points across a landscape. These techniques do not possess the nona-priori knowledge characteristics to extract natural pathways.

Lakes Open land (default factor) Dense woodland Boggy land Sea Visual quality from a point.

In (Kweon In So and Kanade Takeo, 1994) the author described a methodology using contour trees, to extract topographical features such as ridges, peaks, valleys from a contoured DEM. However, in order to extract natural pathways using this technique, we would still need to define pathways in terms of particular topographical structures. This may be relatively easy for a mountain pass, but rather more complex for a pathway that may comprise of several topographical elements such as rivers, rolling hills, and ridges.

The quality of the constraint surface model depends on the weighting, attributed to each of these landscape features. The model could be tailored to suit certain modes of mobility. For example, ifwe assume the availability of boats and ferries for river movement then this may be built into the model. A sedentary farming society may impose different constraints on the landscape which may include territorial boundaries that confine movement. The constraint surface could also be generalised to use subjective attributes such as sacred spaces, forbidden areas or landscape attractors.

A variety of image analysis algorithms, such as skeletization, aim to extract structure from images. These could be applied to DEMs to extract valleys and ridges, but these suffer from similar limitations to topographical feature extraction.

The generation of a good constraint surface is the most important step in the process. The subsequent stages of the algorithm are deterministic. It should be noted that the technique relies on the relative costs between areas and so absolute measures of cost are not needed.

By far the majority of formal methods in landscape studies using GIS, have been applied to hydrology and geology. It was while studying the online documentation within Arcview GIS, relating to hydrological modelling, that a natural pathway finding methodology began to be formulated.

The cost model used in the analyses presented in this paper, uses relative terrain values, or weightings given to each landscape attribute. The effort based cost equation as outlined in (PandolfK B et al. 1977) is not needed, however the relation of cost to slope is still deemed valid for most types of terrain. Some work by Bell, T & Wilson A (submitted for pub 2 ) suggests that the tangent of the slope is a more accurate measure of traversal cost, than slope. However, the natural pathway extraction process utilises relative costs and so the form of the cost function is not crucial. This allows us to derive a simple

Description of "Natural Pathway" Algorithm The path finding process comprises several distinct stages and is an attempt to form a drainage network of a constraint or cost surface. The complete process is illustrated by a flow diagram (figure 1). The procedure and algorithms are identical to the extraction of surface runoff characteristics used in hydrological

6

equation. As the analysis uses a grid based DEM, the equation is described using grid nomenclature.

lookup table defining the most likely direction (Greenlee DD, 1987). The behaviour of the flow direction algorithm, at the edges of the analysis area may be chosen. Cells at the analysis boundary may de deemed to flow inwards rather than to possibly undefined cells outside the area. (SeeArcview Help discussion).

The grid based cost equation used to generate the cost surfaces used in the subsequent analyses, is shown below.

[CJ=

(L( [G]N1[t]+ N1[t]))

+ ((1-2,[t])N1(G+ 1))

This flow direction function is also used as part of the process to create the depressionless surface.

t

Where: [C] [t] N, N1 [G] I[t]

is a grid ofrelative cost values. is a binary grid. Value 1 depicting areas of terrain t. is a terrain value for terrain t. is the default terrain value for open land based on the DEM model. is slope for each cell in the [DEM]. is the sum ofall the terrain binary grids.

The Path Accumulation Grid

This stage creates a grid of accumulated flow each cell, by accumulating the flow from upstream cells using the method described in (Jenson SK and JO Domingue, 1988). This is determined from the flow direction. Output cells with a higher flow accumulation represent areas of the landscape that have a higher probability of being potential pathways.

This particular equation is linear for each terrain factor with a gradient ofG. The relative values ofN, and NI' are important, not their absolute values. If no other constraints are used, the equation reduces to: [C]=GN 1+ N 1

The Path Network

The results of flow accumulation are used to create a path network by applying a threshold value to the path accumulation grid.

It should be noted that further work may be needed to develop better constraint models perhaps incorporating perceptive landscape attributessuch as topographic prominence (Llobera M, 2001).

The resulting, dendritic path networks can be further processed by ordering, identical to stream ordering, using any of the published algorithms (Shreve RL, 1966; Strahler AN, 2001).

The Depressionless Cost Surface Grid

Initial Test Results This is an iterative process and involves the filling of data sinks. A sink is a cell or set of spatially connected cells whose flow direction cannot be assigned one of the eight valid values in a flow direction grid (see below). This can occur when all neighbouring cells are higher than the processing cell, or when two cells flow into each other creating a two-cell loop. As the cell size of the data increases, the number of sinks in a data set often also increases.

Initial tests were applied to a series of synthetic cost surfaces to test the behaviour of the technique. These clearly showed the validity of the methodology and its limitations. In order to provide a real test of the "natural pathway" extraction process, a digital terrain model of an area of approximately 430km2, including sections of the Chiltern Hills, was analysed. The DEM was extracted from the UK Land-Form PANORAMA data set, and consists of height values at each intersection of a 50m horizontal grid, the values have been mathematically interpolated from the contours on 1:50000 maps. DEM height accuracy is no greater than one half of the vertical interval of the source contour data, ( 10m). The water features were extracted from the UK STRATEGI dataset, these are at a scale of 1:250000.

To create an accurate representation of flow direction and therefore accumulated flow, it is best to use a data set that is free of sinks. This ensures that sinks created by sampling errors are filled, and paths do not drain into artificial sinks. Naturally occurring cost sinks may also occur. For example, a steep circular conical hill with a flat top would be a morphological sink which if filled would result in a circular pathway at its bottom but none on its top, which could be a valid pathway. Sinks are filled to the lowest cost "pour point".

The area is also crossed by a long distance footpath, the Ridgeway.

The Path Direction Grid

The results of the tests are illustrated in figs 2-3.

One of the keys to deriving the path network is the ability to determine the direction of flow from every cell in the grid. This is calculated using a flow direction algorithm (Jenson S Kand J 0 Domingue, 1988) which outputs a grid showing the direction of flow out of each cell. The direction of flow is determined by finding the direction of best gain, that is, to the lowest cost value. If the gradient is the same to all the cells then the neighbourhood is enlarged until the steepest gradient is found. If a cell has the same change in cost value in multiple directions and is not part of a sink, the flow direction is assigned with a

Fig. 2 is a coloured representation of a DEM for an area that covers part of the Ridgeway long distance footpath. The graticule shows the National Grid Co-ordinates. The modern route is highlighted in black. The path network, in yellow, was created from thresholding the path accumulation grid to 1000 cells. A terrain value of 10 and 1 was used for water and land respectively. A purely visual inspection shows that the approach is reasonable. It can be seen that paths in hilly terrain generally

7

follow ridges. It can also be seen that the methodology produces many potential path networks in relatively flat plains. This is to be expected in the absence of any constraining terrain. It is probable that in prehistory, swathes of dense woodland or wetlands existed in these flat plains near to rivers. These terrain types would be areas of high cost. An estimation of the extent and nature of woodland and any other constraints would therefore be important in order to gain a better understanding of natural pathways in otherwise flat unconstrained areas.

consequence of grid based schemes, which can be improved by artificially thickening the rivers. This limitation should only affect the accuracy close to rivers. The path direction grid has only eight possible values, which is simplistic, although the digital grid based nature of the DEM defines this limitation. A possible approach to improve this is by incorporating the parsimonious techniques described in (Douglas David, 1994) as a refinement to the flow direction algorithm.

The very high cost attributed to water produces paths that avoid crossing rivers. This model is therefore too simplistic as some of the rivers are slow moving lower sections, (usually navigable), but serves to test the behaviour of the technique and cost surface equation.

The technique does not attempt to connect natural pathways together. The located paths will generally avoid high cost areas. A long distance path may comprise segments of natural pathways combined by the crossing of relatively small, high cost areas.

No account of other terrain costs has been taken. The modem Ridgeway path will also follow routes of low cost and deviate from areas of high cost, which became high in modem times, such as defined rights of way, agricultural land. The modem paths cross modem river bridges, which are low cost, artificial elements of the landscape.

The accuracy of paths close to the edges of the area being analysed depends on typical boundary effects. For the tests described in this paper, the analysis area was a subset of a larger DEM. For DEMs that include sections of coastline, the sea needs to be correctly cost modelled, or the coastline must be defined as a boundary, otherwise the sea may act as a pathway drain.

Fig. 3 shows the same area as fig. 2, but in this case, a terrain value of zero for water was used to produce pathways that ignore the river constraints. Note that this produces different results for paths in the relatively flat areas, as these have numerous nvers. These paths frequently cross the rivers as expected.

The cost grid is currently isotropic and this characteristic may need further investigation.

Natural Pathways around Stonehenge: An Initial investigation into their Relationship to Prehistoric Ritual Sites

Discussion on Methodology There are several factors limiting the accuracy of the paths, created by using the "natural pathway" derivation technique.

With due consideration to the limitations of the natural pathway extraction technique, an attempt was made to investigate potential patterns and relationships between the natural pathways around Stonehenge, and the surrounding ritual monuments.

The resolution of the DEM and the accuracy of the elevation data are a major limiting factor. Coarse cells can provide only average height data, and so may hide small areas of high cost which would affect path location. This will effect the minor paths in a network more than the high thresholded segments. The effect of cell size is lessened in areas of gentle topography.

An area of approximately 30km2 around the site of Stonehenge was analysed using the path finding method. The DEM used was the Land-Form PANORAMA data from the Ordnance Survey, as described previously. This was the only constraint data utilised in this analysis due to the lack of appropriate resolution river data. The available river data set at 1:250,000, had evident misregistration with the DEM data. It was considered reasonable to leave out the river constraints in this area as there are virtually no minor rivers around Stonehenge. Instead, the larger area is notable by the presence of a few navigable rivers including the Avon.

The path finding method is very sensitive to the input constraints. The accuracy of the path networks depends on a realistic model of terrain costs. This cost model could be subjective, and limited by a lack of appropriate data. As mentioned previously, in the absence of other constraints, the method is robust to the slope function used for cost. In areas of land with visible topography, the technique will locate natural pathways as defined by the absence of any other unknown constraints. However, in flat terrain other constraints would be needed in order to provide more meaningful pathways.

An archaeological site database, provided by English Heritage, was used to investigate potential relationships. The archaeological data included information on; Long Barrows, Round Barrows, Henges, Causewayed Enclosures and Ring Ditches. Basic data on the cursus monuments and the Stonehenge Palisade was also included.

The river features need to be converted to a grid with a cell size equal to the DEM cell size. The PANO RAMA data used in the previous analysis has a cell size of 50m, producing, in many UK cases, an unrealistically large river. If 2 river grid cells, join at the corners, the river grid may be incorrectly crossed by pathways across the data gap. This is an unavoidable

The results of the analysis are shown in figs 4 and 5. Fig. 4 shows a coloured representation of a DEM for an area centred on Stonehenge. The path network denoted by a thick

8

yellow line, was created from thresholding the path accumulation grid to 1000 cells. This is overlaid by a thin yellow line, which is the result of thresholding the path grid to 100 cells. A terrain value of one was used for land.

Woodhenge and just beyond, joining the round barrows nearby. Some of these initial observations will form the basis of a further, in depth study, of the Stonehenge area. Further analyses will utilise the temporal dimension to investigate the existence and nature of any chronological patterns.

As discussed previously, It is possible that in prehistory, swathes of dense woodland existed in the valleys.

Future Work Fig. 4 shows a small section of the analysis area centred on Stonehenge. The path network, denoted by a thick yellow line, was created from thresholding the path accumulation grid to 1000 cells. This is overlaid by two other yellow lines, which are the result of thresholding the path grid to 100 and 20 cells.

As mentioned previously, one of the main limitations to the accuracy of the path network is the quality of the constraint model. Therefore, more research will be directed to improve the model with the addition of other landscape and perceptive components. These are to include;-

Discussion of Results

Dense woodland River flow direction River type (navigable-low cost) mountain river (high cost) Wetlands or Bogs. Sea Visual quality: low cost attributed to cells with rich visual affordances.

Some possible relationships between the prehistoric ritual sites and the natural pathways can be speculated from a purely visual analysis of the images. One of these is the possible correlation of round barrows near the paths. To test if this potential relationship was non-random, a Kolmogorov-Smirnov analysis was performed. For the 874 round barrows in the analysis area, the difference between the cumulative area distribution and the cumulative site frequencies at several distances from the pathways was calculated. The results are shown in Table 1. The figures are based upon pathways which have a been thresholded at 20 cells.

The path accumulation stage can also be combined with a weighting grid. Knowledge of settlement sites could be used to produce a weighting grid to enhance the path network, although this information is scant for prehistoric Britain. The relationship between threshold level and KolmogorovSmirnov test results will be investigated in more detail.

The difference between the two distributions was found to be significant at below the 5% level. Although this result provides evidence of some relationship, it does not provide causal information. In addition the test does not allow for any information on the structural distribution of the round barrows. For example, from a purely visual analysis, some of the round barrow clusters seem to be congregated at pathway cul-de-sacs.

The pathway approach will be attempted with a much larger DEM of the whole UK. Potential relationships between these pathways and an archaeological database of over 20000 ritual Neolithic and Bronze age sites, will be explored.

Conclusions It could be conjectured that many of the long barrows seem to

have a similar alignment to major paths. However it has been suggested that many long barrows are aligned along ridges and so this may be the cause of the visual pattern. Since slope was the only constraint used for this analysis, the natural pathway technique will locate ridges and valleys.

From the outset, it has been our intention to be able to study the potential correlation between natural pathways through a landscape, and prehistoric ritual sites in the UK. This led to a need to develop a technique to find such pathways, independently of site location. A method to extract "natural pathways" was formulated. The nature of the techniques, based on terrain cost, produces a network of potential natural pathways. These path networks do not exclude the possibility of other, perhaps minor paths, existing between path networks or across short stretches of high cost terrain. The higher order pathways are probably more reliable as indicators of major potential thoroughfares.

An initial look at the results might suggest that the henge monuments seem to be closer to the major pathways. A major pathway heads off to the Northwest from the Stonehenge area in the direction of Robin Hoods Ball, another major prehistoric site. At Stonehenge, a spur off the major pathway to the Northwest, seems to lead to Stonehenge going through the palisade. It is known that a gap exists in the palisade structure, perhaps a gateway for the pathway?

The results of the initial applications suggest that research into the relationships between archaeological sites and "natural pathways" is worth developing further.

The Stonehenge Avenue, interestingly has only a vague relationship to the paths, but its changes of direction are echoed in the major path some 300m to the North.

The natural pathways process, provides archaeologists with a useful tool to investigate the landscape, that may improve the interpretation of site or artifact distributions. It may also allow archaeologists to provide structural meaning to inter-site lands-

It is interesting to note that a small path starts at Durrington Walls, leaving near the known SE entrance, and goes to

9

cape studies. Ingold, T. 1993. The Temporality of the Landscape. In: WORLD ARCHAEOLOGY 25, 152-174.

Furthermore, the approach is generic and could be used for exploring any constraint surface model.

Jens on, S. K. and Domingue,]. D. 1988. Extracting Topographic Structure from Digital Elevation Data for Geographic Information System Analysis. In: Photogrammetric Engineering and Remote Sensing 54, 1593-1600.

Acknowledgements I wish to acknowledge English Heritage for providing the archaeological data. The DEM data and river data have come from the Ordnance Survey via EDINA's DIGIMAP service. I would like to thank Proffessor Richard Bradley and Dr Vince Gaffney for their enthusiastic support and the University of Birmingham for supporting my participation at CAA2001. My thanks also go to Garry Lock, Tyler Bell and the many other people who commented on the presentation at CAA2001.

Kweon In So and Kanade Takeo. 1994. Extracting Topographical terrain features from elevation maps. In: Computer Vision, Graphics and Image Processing:Image Understanding 59, 171-182. Llobera, M. 1996. Exploring the topography of mind: GIS, social space and archaeology. In: ANTIQUITY 70, 612-622. Loveday, R. 1998. Double Entrance Henges-Routes to the Past? In: Gibson, A. and Simpson, D. (Eds.) Prehistoric Ritual and Religion, pp. 14-31. Sutton]

End notes 1Bell, T. and Lock, G. Information discussed by private communications with the author. 2 Bell, T. and Wilson, A (submitted for pub) "Tracking the Samnites: landscape and communications routes in the Sangro Valley, Italy". Extract received by private communication from Bell, T. 3A recent paper by Marcos Llobera in "Beyond the Map, 2000" edited by Lock, G. explained to the author by Lock, G by private communication at CAA2001.

Llobera, M. 2001. Building past landscape perception with GIS: Understanding topographic prominence. In: Journal of Archaeological Science 28 (9). Pandolf, K.B., Givoni, B. and Goldman, R.F. 1977. Predicting energy expenditure with loads while standing or walking very slowly. In: Journal of Applied Physiology 43, 577-581. Parcero Oubina, C., Criado Boado, F. and Santos Estevez, M. 1998. Rewriting landscape: incorporating sacred landscapes into cultural traditions. In: WORLD ARCHAEOLOGY 30, 159-176.

References

Shreve, R.L. 1966. Statistical Law of Stream Number. In: Journal of Geology 74, 17-37.

Bradley, R., Boado, F.C. and Valcarce, R.F. 1994. Rock Art Research As Landscape Archaeology - A Pilot-Study in Galicia, North-West Spain. In: WORLD ARCHAEOLOGY 25, 374-390. Douglas, D. 1994. Least Cost Path in GIS using accumulated cost surface and slope lines. In: Cartographica 31, 37-51.

Strahler, A.N. 2001. Quantitative Analysis of Watershed Geomorphology. In: Transactions of the American Geophysical Union 8, 913-920.

Greenlee, D. D. 1987. Raster and Vector Processing for Scanned Linework. In: Photogrammetric Engineering and Remote Sensing 53, 1383-1387.

Tilley, C. 1996. The powers of rocks: Topography and monument construction on Bodmin Moor. In: WORLD ARCHAEOLOGY 28, 161-176.

Tables

distance range to pathway (m) 0-50 50-100 100-150 150-200

cumulative area ofland (m2)

area (percentage of total)

358369984 526497488 699492496 792272496

0.416974 0.612595 0.813879 0.921831

Total

859455000

Dmax

Table 1.

cumulative number of round barrows 354 579 745 828

round barrows (percentage of total) 0.405034 0.662471 0.852403 0.947368

874

is Greatest cumulative difference

0.049877

Critical value for 5% level is 1.36/✓ 874

0.046003

Critical value for 1% level is 1.63/✓ 874

0.055136

10

difference between cumulative distributions -0.01194 0.049877 0.038524 0.025537

Figures

L

COST SIJRFACE EQIJA'IJON

Cost Surface Grid

JifLL SINKS

Depressionless Cost Surface Grid

Figure 2: Natural Pathways in the Ridgeway area with river constraints !J•ct, M-1cl•-1,

~uuuri;,

Ur;1uu~it,

c,f s,dn-1y

C:r-1-1t-1dby Ll-1w-1llyn, l'.ich ■ rd Sydri-1y, Hi~toric, Not y•t r-1t•d

Chinese

Biographical

Database

maintained

Published by LE!w1s-Clark State Credted by Levine, Marilyn A. China, Not yE!t rdted

CollegE!

National

Database

Archaeological

by Marilyn

Figure 9: Sydney TimeMap dual screen interface Macromedia Director wrapper

Levine

.Jo)JIJ

(NADB)

Reports

·=I

[USA]

Published by US Ndt1ondl Park Se,rvice Cee"t"d by Childs, T""Y USA, Pre,hidoric, Notye,trated

r

45

.:;;_, r 48

Household

data for The Rocks 1865

Published by NA (unpublished thesis by Wdyne Historic-al Arch-aE!ology, University of SydnE!y) C,e"t"d by Mullen, Wayn" NotyE!trated

Stein's

photographs

of archaeological

Mullen 1993,

held by the Department

of Prehistoric

r

dnd

El Afl,•r!1hul ....

,

!.nrr.

;; Cll•ll ♦••t1PDII•• r El ■ur•itun 1:1••1

sites

early

m,f'•,\;~

1900s

P' El Jm;ig;;,vlf.wpointr.

Published by ACL c,.,,.t,.d by Pdsul Simon China, Not ye,t r-ated

'

AfrQcW :i;hQW~ dlrt1"1.tlon

P'El\/ii!itll.Ylf.wpointr.

'

AfrQcW :i;hQW~ dlrt1"1.tlon

(:llEER.~i:•n.,t:·;

Figure 7. ECAI clearinghouse search results showing web sites, datasets and passworded datasets. Datasets are included in the map result by checking the boxes

r

~-~•tl"rn

••"t' ,1.llllln,

IEIMQ'd•..rnffi"E"'l't1 ~

(.1EJMQ'd'l'rni;cQ,nd!nf!" /\/

Figure 10. A historical map overlaid with modern coastline, image viewpoints and directions

Figure 8. The Sydney TimeMap kiosk application at the Museum of Sydney

77

Sydney TimeMap Assets Listing

Sydney TimeMap Asset 36

Figure 11. Web-based access to the digital image database showing thumbnail views and a higher resolution image generated on the fly

Figure 13. An 1844 map showing the effect of scale error across Sydney Cove (centre)

Figure 12. Webpages generated from the Sydney TimeMap kiosk application by the thumbnail template (above) and by the single item template (below).

Figure 14. Map of Sydney presented as a historical document on a 'map table' display (detail of map shows use of drop-shadow to emphasise the map as a historical document)

78

The answer is blowin' in the wind. Research desires and data possibilities. Hans Kamermans Faculty of Archaeology Leiden University P.O. Box 9515 NL 2300 RA Leiden Phone: +31 71 527 2385 - Fax: +31 71 527 2429 - E-mail: [email protected]

Abstract: Thispaper was supposed to be called "Buried by the wind. Regional anazvsis in Palaeolithic and Mesolithic Arclzaeolo6ry in the Southern Netherland~". It was supposed to examines tlze influence of loess deposits on the distribution of Palaeolithic and Mesolithic find spots in Limburg, the most southern region of the Netherlands, using a Geographic fnfcmnation System (GIS). ft was supposed to demonstrate that the archaeological visibility of Palaeolithicjind spots in that part of the Netherlands is greatly hindered by the Upper Pleistocene loess cover of this part of the country. ft was supposed to demonstrate also that the visibility of l'vfesolithicfind spots is not influenced by that geological phenomenon. And it was supposed to investigate if these differences in tlze distribution of/ind ,1pots,and the concluded differences in land use, are a result of geological processes. But the quality of the data set did not make any of this possible. Key words: GfS, Palaeolithic Archaeology, land use.

Introduction There is no problem with the theoretical background, but what about the research question?

The following research is part of an ongoing project about the Palaeolithic of Limburg, the southern pmi of the Netherlands (Kamermans & Rensink 1999). This research tries to analyse and interpret Palaeolithic and Mesolithic find spots from the Dutch loess area from a landscape perspective. One of the problems is how to evaluate cmTectly areas without find spots. Does this point to a selective use of specific zones by huntergatherers, or are these 'empty' areas a result of geological or recovery processes? In more general terms: to what degree is the observed distribution of stone artefacts across the area representative for the use of the landscape by prehistoric hunter-gatherers?

Research question There is an ongoing debate about differences in subsistence strategies during the Middle and Upper Palaeolithic. What is the difference in subsistence strategy between the Neanderthals, or the Ancients as Stringer and Gamble (1993) call all pre-modern humans, and the Modems? For instance, did Neanderthals hunt? If yes, were they general or specialised hunters? Most researchers see the change in subsistence happening between Ancients and Moderns, and considered it as pmi of the so-called Middle to Upper Palaeolithic "transition". Others see changes during the Middle Palaeolithic (c. f. Stiner 1994, Kuhn 1995). Differences in subsistence, means differences in human behaviour and should leave a difference in spatial patterning of the material culture in the landscape. The problem is there. Can we solve it in Limburg?

This particular part of the research stmied with the following question: ·'We assume a different economy and a different land use between the Palaeolithic and the Mesolithic period, or at least between the Middle Palaeolithic and the Mesolithic. Can one see this in the distribution of sites in the landscape in South Limburg or is this pattern influenced by the geology? The Middle Palaeolithic sites were covered by Ioess during the Pleniglacial and the later Mesolithic sites were not. Are we looking at a difference in distribution or are we simply looking at moreorless the same pattern obscured by the deposition of loess. What if we use GIS and throw a loess cover over the Mesolithic sites. Would the same pattern as for the Palaeolithic emerge?"

The data. set

For good research you need a good theoretical background, a good research question, a good dataset, and good tools to research your question. It looked as if all this was available fi.1r our Limburg case study.

79

Limburg is archaeologically speaking a very well researched area. Not only for archaeology in general but especially for the Palaeolithic period. Roebroeks excavated the Belvedere quarry with the oldest site from the Netherlands, 250.000 years old (Roebroeks 1988). Rensink studied the Magdalenian sites in Lin1burg and surroundings (Rensink 1993). Amateur archaeologists are very active in that part of our country and for GIS applications Limburg is perfect; it is one of the few places in the Netherlands with some form of relief and has a

a surface find or an excavation that will help. It does, but as I said we used a slightly updated ARCHIS dataset and we only included surface sites, so the problem remains.

great variety in landscape. But for a GIS application you need both maps and archaeological data. For this research we used two geological maps, one soil map, one geomorphological map, two maps with the distribution of the loess, and a slope map and a slope aspect map from parts of the area. For the archaeology we used a slightly updated dataset from ARCHIS, the national Dutch sites and monuments record.

A similar analysis on data collected by the Agro Pontino Project (Voorrips et al. 1991) utilized the field database and used maps only for illustrations (Kamermans 2000). The data that was registered in the field included not only the parent material, but also the soil type, the slope angle and the slope aspect, the geomorphology, and dozen's of other things (Voorrips et al. 1989). We could perform the same kind of analysis for our Limburg research but is this kind of information available in the Dutch national database ARCHIS?

Figure l shows the distribution ofloess and the location ofthe Middle Palaeolithic and Mesolithic sites. The loess map indicates areas with a loess cover of more than l meter thick, loam on slopes, stream sediments and areas with no loess. The loess cover was deposited during the Pleniglacial, lets say from 55.000 until 13.000 year ago. Some of the Upper Palaeolithic sites (from 35.000 until 10.000) date from during this period of deposition and some do not. Including these Upper Palaeolithic sites will complicate matters so we leave that period out of our analysis. With a visual inspection we do not see a difference in distribution.

There is a lot of information in the database. Administrative information about the locality, information about the landscape and information about the site and the archaeological material. There is information about the geomorphology, the geology, the texture of the soil and the situation in the landscape. But there are two problems. It is not always clear if this information is collected on site or taken from a map and the information is not always there. On the form people have to fill in when they report a new find, a suggestion is printed next to the geomorphology field, which is to take this information from the geomorphological map. In the case of the current research therefore, AR CHIS did not help very much. This is the second problem with the data set.

Table l gives the relation between Middle Palaeolithic sites and the various deposits and shows us as many sites on the loess as expected and twice as many on the slopes as expected. Table 2, the relation between Mesolithic sites and the various deposits, gives more or less the same picture. On the loess are as many sites as expected, on the slopes twice as many.

Conclusions The Middle Palaeolithic sites on the loess are of course a problem. They date from before the loess and, for that reason, cannot be on top of the loess. The map (figure 2) confirms this.

With a very straightforward application, like the one sketched above, we encountered two major problems: the digital maps and the archaeological data set. Will these problems be solved in the near future?

This shows us the first problem with our data set. The maps are apparently not accurate enough although the resolution is 25 meters. The scale of the original maps is l : 100.000.

A new development with digital maps in the Nether lands is the AHN (Actueel Hoogtebestand Nederland), a new digital elevation database from the whole of the country made with laser altimetry (http://www.minvenw.nl/rws/mdi/ geoloket/ index I.html). It has a density from 1 point per 16 m 2 • This 3D dataset is accurate enough to solve in the near future the problem of the relation between slope and Ioess coverage.

Another option is to use the slope map. There is a relation between loess and slope. The loess cover is still present on relatively flat surfaces. We have a slope map from part of the area, the Central Plateau. Maybe the slope map will show that the Middle Palaeolithic sites on loess are in reality lying on steep, eroded slopes (figure 3). The table (3) gives more information. One fourth of the number of sites that could be expected on the basis of chance lie on the fiat area and more sites than expected lie on the slopes.

But the digital geological map, soil map and loess map will not become more accurate than the original maps. So these maps are not accurate enough to rely on the information. To do the analysis on the basis of a database we will need an awful lot of data in ARCHIS. We need information gathered on site about the geology, geomorpholo 6,y, soils (like parent material, texture, soil type}, drainage, slope angle, slope etc, etc. Only then we can do our analysis independent of our maps.

For the Mesolithic the number of sites on the fiat surface is as expected and on the slopes more than expected (table 4). It looks as if the loess does influence the pattern. But there are still Middle Palaeolithic sites lying on the fiat surface. So also the slope map is not accurate enough.

The ROB is planning a new version of AR CIUS. It will have three layers of information. One fix one for scientists and one for lay people. Every layer would have their own kind of information. From the documentation (Van Capelleveen et al. 2000) it becomes clear that the expert layer will contain administrative information to manage the site. The form to a site into ARCHIS will stay more or less the same, so it looks as if not all of the information required for a good

We know that most maps are not accurate ( soil maps for instance have an average accuracy of 70% (Kamem1ans & Rensink 1999). So much for GIS as a tool to solve all your problems. Bui is it possible to do the analysis without maps? Is there enough information in the database? If the database tells us if the site is

80

regional analysis will be there.

archaeology. A case study from the southern Netherlands. In: L. Dingwall, S.C. Exon, V.L. Gaffney, S. Laflin & P.M. van Leusen (eds). Archaeology in the Age of Internet - C4A97. Computer Applications and Quantitative A1ethods in Archaeology. BAR International Series 750: 81 & CD-ROM (ca. 30 p.).

The problem is highly relevant. One of the most important products of the State Service at the moment is the IKA W, the indicative map of archaeological values (Deeben et al. 1997). These maps are a result of predictive modelling on the basis of AR CHIS and digital maps of the physical environment. There has been a lot of criticism directed towards the first t\vo generations of these maps (c.f. Verhagen et al. 2000). The new ARCHIS will play an important role in the production of future predictive maps. In my opinion it is impossible to predict site location for CRM purposes without a very detailed archaeological database. The only other solution is to rely on the unreliable digital maps.

Kame1mans, H. 2000. Land evaluation as predictive modelling: a deductive approach. In: G. Lock (ed.). Beyond the Map. Archaeology and Spatial Technologies. NATO Sciences Series. IOS Press: Amsterdam. 124-146. Kuhn, S.L. 1995. Mousterian Lithic Technology.An Ecological Perspective. Princeton University Press: Princeton, New Jersey.

With the general available data it was impossible to answer the simple question formulated above. The problem remains, so in the end the answe1; myfriend, is blowin 'in the wind.

Rensink, E. 1993. Moving into the North: Magdalenian Occupation and Exploitation of the Loess Landscape of Northwestern Europe. Unpublished PhD thesis, Leiden.

Acknowledgements

Roebroeks, W. 1988. From Find Scatters to Early Hominid Behaviour. A study of Middle Palaeolithic River Side Settlement at Maastricht-Belvedere (The Netherlands). Analecta Praehistorica Leidensia 21

I would like to thank my colleague Professor Wil Roebroeks (Faculty of Archaeology, Leiden University) for the research question I could not answer and Eelco Rensink (Dutch State Archaeological Service) for the collaboration in the Limburg project. Both would have been co-author's of this paper had I answered the research question. I am very grateful to Martijn van Leusen (Groningen Institute of Archaeology, Groningen University) for making all his digital maps of South Limburg available to me.

Stiner, M.C. 1994. Honor Among Thieves. A Zooarchaeological Study ofNeandertal Ecology. Princeton University Press: Princeton, New Jersey. Stringer, C. & C. Gamble. 1993. In Search of the Neanderthals. Solving the Puzzle of Human Origins. Thames and Hudson: London. Verhagen, P., M. Wansleeben & M. van Leusen. 2000. Predictive Modelling in the Nether lands. The prediction of archaeological values in Cultural Resource Management and academic research. In: Harl, 0. (ed.): Archiieologie und Computer 1999. Forschungsgeselschaft Wiener Stadtarchiieologie 4: 66-82.

References Capelleveen, E.J. van, E.E.A.M.C. Kuis & J.C. Meerkerk. 2000. Rijksdienst voor het Oudheidkundig Bodemonder::oek. Programma van Eisen ten behoeve van Europese aanbesteding vemieuwing Archis. Twynstra Gudde, Amersfoort.

Voorrips, A., S.H. Loving & H. Kamermans. 1989. Infom1ation Science in Archaeological Survey. in: T. Hackens & U. Miller (eds). Geology and Palaeoecology/or Archaeologists. Pact 24: 189-210.

Deeben, J., D. Hallewas, J. Kolen & R. Wiemer. 1997. Beyond the crystal ball: predictive modelling as a tool in archaeological heritage management and occupation history. In: Willems, W., H. Kars & D. Hallewas (eds.): Archaeological Heritage Management in the Netherlands. Fifty Years State Service for Archaeological Investigations. ROB, Amersfoort, pp. 76-118.

Voorrips, A., S.H. Loving & H. Kamermans (eds.). 1991. The Agro Pontino Survey Project. A1ethodsand preliminary results. Studies in Prae- en Protohistorie 6. Amsterdam. 131 p.

Kamermans, H. & E. Rensink. 1999. GIS in Palaeolithic

81

Tables unit

area

loss cover loam on slopes stream sediments no loss

0 sites E sites

perc

346.47125 163.47375 53.19625 129.334375

50.03 23.61 7.68 18.68

35 37 0 6

39.03 18.41 5.99 14.57

692.475625

100.00

78

78.00

Table 1. The relation between Middle Palaeolithic sites and the various deposits. Area is in km 2, 0 sites is observed sites and E sites is expected sites unit

area

loss cover loam on slopes stream sediments no loss

0 sites E sites

perc

346.47125 163.47375 53.19625 129.334375

50.03 23.61 7.68 18.68

19 24 1 1

22.52 10.62 3.46 8.40

692.475625 100.00 45 45.00 Table 2 the relation between A1esolithic sites and the various deposits unit

low medium high

area

perc

0 sites

E sites

61.82 8.60 8.00

78.83 10.97 10.20

2 6 2

7.88 1.10 1.02

78.42

100.00

10

10.00

Table 3. the relation between Middle Palaeolithic sites and slope class. unit

low medium high

area

61.82 8.60 8.00

perc

78.83 10.97 10.20

0 sites

E sites

5 3 0

6.31 0.88 0.82

78.42 100.00 8.00 8 Table 4. the relation between Afesolithic sites and slope class.

82

Palaeolithic and Mesolithic sites in South Limburg 11111111111 loess

Middle Palaeolithic sites in South Limburg

cover

11111111111 loess

cover

c::::Jloam on slopes c::::Jstream sediments c::::Jno loess

c::::Jloam on slopes c::::Jstream sediments c::::Jno loess

1[11Dfil

1[11Dfil

Figure 1. The distribution ofloess and the location of the A1iddle Palaeolithic (white dots) and Mesolithic (red dots) sites.

Figure 2. Ihe distribution ofloess and the location of the Middle Palaeolithic sites.

Middle Palaeolithic sites Central Plateau low relief mediumrelief highreliel

Figure 3. Middle Palaeolithic sites on the slope map.fi'om the Central Plateau.

83

A hundred years of lake contour fluctuation in the Hamun-i Helmand: A GIS based system for the study and the recovery of archaeological information in the Iranian Sistan (1899-1999) Sabatino Laurenza University of Bologna Dept. ofArchacology, Italy via G. Spontini 9/A 00013 Mcntana (Roma), Italy Phone: +39 06 9091843 - Mob: +39 347 5250441 - E-mail: [email protected]

Sophie Pornet-Laurenza University of Bologna Dept. of Archaeology, Italy via G. Spontini 9/A 00013 Mcntana (Roma), Italy

Abstract: The fragmentation of the contextual set of all kinds of information, which has not been preserved intact during the years, has always been a big limitfor archaeological research. Reconstruction is now realisable tt•ith a stochastic process, which gives a complete picture of the pertinent data by increasing approximations, because it is able to join together information and contexts otherwise scattered. The recording, display, management and storage of data by computers, gave us the possibility to do complex analysis, which was previously impossible. in this framework GIS are a set of instruments able to push archaeological data analysisfurther. l(we consider the spatial nature of archaeological data - the recording of the precise location o(artefacts, ecofacts, architectural or stratigraphical units and sites - the GJS allows the archaeologist to match different types of geographic data (hydrography, soils, elevations, ...) in order to produce wide-scale analyses. Key words: Historical cartography, Iranian Sistan, Lake.fluctuation, GIS~Remote Sensing.

Introduction

The Iranian Sistan

Apart from a few and very exceptional cases, the archaeological record all over the world is too fragmentary, to return a direct vision of ancient lives and landscapes. The context is usually lost and it is left to the skills of specialists "to piece together the fragments of the past". Generations since World War U have been blessed by computer science, with an increasing capacity to combine vast archives of data. Reconstruction has been made possible, more recently, by so-called "stochastic procedures" which allow combining of very large numbers of operations, modeling the scattered fragments into true pictures of lost realities. As more data is introduced with different theoretical pcrspcctivs, models can be constantly updated.

The Hilmand Delta in Sistan represents a clear case of instability evidently connected to the fact that the basin is sharply limited to the west by a North-South barrier made by the front of the Palang Mountains and their piedmont colluvial fans of gravel sediments. The first noticeable difference between the Hilmand and the other systems, such as the Murghab or the Nile, is the fact that the densest settlement areas and the capital centres have always been, since the end of the 4th millennium BC, at the end of the delta system, to connect with the river as well as the huge terminal lake. The rapid build-up of sediments compels the waters to change their course several times, while a circular anticlockwise sequence of descending lacustrinc basins connected with each other by spillways, like the Shcclag rud, drains overflows to the southernmost sections of the basin (fig. 1). Satellite imagery of southern Sistan indicates some six or seven overlapping delta fans, all dated by the archaeological sites to later historical times, spanning less than a thousand years in Islamic times. Even more unstable have been the southern limits of the Hamun, the largest and most perennial of the ter-

Geographical Information Systems (GIS) at present arc the main instrument available from computer science to reconstruct past landscapes and human populations in their historical fluctuations, by combining data and information from maps, earth sciences, surveys and excavations.

85

minal lakes. What is left of this ancient system, mostly dating to protohistorical times, is reduced to thinyardangs, columnar sedimentary residues shaped by the dominant winds. Third millennium BC soils, ofi:en black from organic content, have been detected at the top of these yardangs about two meters above the present level. The combined action of the expanding lake and wind erosion has lowered most of the ancient surfaces, destroying all residual evidence of the Bronze Age landscapes.

he describes as an island. Thirty years later, Major Lovett (Rawlinson, H. R. 1873.) crosses the dried lake by foot to reach the same point (fig.3)

The site Discoveries during the last thirty years in Central Asia (particularly in the Iranian area) brought to light the antiquity of the proto-urban process in that region of the world far away from the traditional civilizations (Egypt, Mesopotamia, Harappa).

The analysis of the Aeolian deflation, as a process of concentration of the objects through the effect of the lowering of the deposit layers needs to re-evaluate the surface documentation that precedes different strategies of research. Wind and water are tireless excavators, that follow scare but precise mles (fig.2 ). During the 70's,American researchers like C. Redman and P.J. Watson (Redman, C. R., Watson, P. J. 1971) worked on the information drawn from the dispersion of stone artefacts on the surface of prehistorical sites in oriental Anatolia. Between 1966 and 1970, measures of the Aeolian-raining erosion rates in the Oaxaca valleys in Mexico and in the Deh Luran in Iran were made and allow Kirkby and Kirkby (Kirkby, A., Kirkby M. J. 1976.) to elaborate models of geo-morphological alteration. The archaeological team of SiS, convinced by the interest that this kind of study would present in the Sistan area, couldn't apply these methods of research, as the site of SiS was much bigger than the areas studied by the Americans and the wind and water freaks are much more powerful (Tosi, M. 1985). In 1972, they applied systematic surface survey to the areas of the site where the handicraft-quarters for lapis lazuli at SiS were discovered.

The state fi:)mrntion process comes for the first time, between 3400 and 3000 BC in the Nile valley, in Mesopotamia, in the plateau regions oflran, Turkmenia, and in the Indus valley thanks to the exceptional demographic increase, to the densification of settlements along the hydrographic axes, to the exploitation of the mineral resources organized following work specialization, and to the hierarchical establishment of the society. Between 3200 and 2900 BC hundreds oflittle settlements are abandoned, creating fix the first time a precise difference between center and periphery. In such period urban centers and as Hissar, Tureng Tepe, Namazga Tepe and Mundigak, which, together with Shahr-i Sokhta (hereafter S.iS.) will represent the "Helmand civilization". The principal settlement of Sistan, S.iS., which represents the main center of the region during the protohistorical period, is founded on the top of the terrace, at the end of the fourth millennium.

Historica] cartography

The excavations of the Italian archaeological expedition of IsIAO (ex IsMEO) recognized during ten years of research at least 11 cultural phases (also called "structural phases") of development of the city. S.iS. represents a trade center for local and extra-regional products, but also a high-specialized place of artifact production, with entire urban and extra-wall areas dedicated to these activities.

The pivotal position of Sistan, among Afghanistan, Baluchistan and Iran, was critical in the power struggle between the Russian and the British empires fix the region of Middle Asia during most of the 19th century. This conflict has ultimately determined the present political division of the region

II. C. Rawlinson (Rawlinson, H. R. 1873.) who visited Sistan in 1872, writes that "the comparative geography of Seistan is rendered especially difficult by the constant changes in the lower course of the Helmend, changes which are common to the deltas of most rivers; but which occur with all the greater frequency when owing to the very slight difference of level, there is no defined basin to receive the waters that are discharged". It is also useful to underline the difficulties, others British officers found in defining the course of the main river - the Hilmand from its secondary branches, and to make the difference between artificial canals dug for irrigation purposes and natural rivers beds.

The secondary settlements (between 2 and 6 ha.) almost all belong to the protostate phase (from 2500 BC), characterized on one side by cultural development, by technological progress and by an enlargement of the settled areas; on the other side, begins the development of a cultural preservation, due mainly to the shift of the traditional commercial and cultural axes. This process increases during the second millennium for the disappearance of the urban structures in the Iranian area. The crisis ofS.i.S. in the frame of the Sistan protostate system began around the beginning of the second millennium when the town becomes smaller (from 120 ha. to 5-6 ha.) (fig. 4).

Before the work of the 1899-1902 Afghan-Iranian Border Commission, under the Head of Col. M. Mac Mahon who produced the first complete and detailed maps of the area and described precisely the circular movement of the water of the river and the lakes, we can just have an idea of the variability reading the different papers of the British officers, in particular telling about their visit to the Koh-i Kwaja mount. In 1842, Captain Conolly (Rawlinson, H. R. 1873.) needs a small reed-boat to reach what

Modern Cartography and Satemte Imagery The opportunity of using historical and modern maps, and especially satellite photos now available, in such a way as to understand the interactions between man and landscape through the reconstruction of environmental changes and their consequences on the settlement in this region, can be considered an extension of the purposes ofthe Italian Mission in Sistan,

86

helped by new technologies. The present work aims to enlarge the use of these new tools looking "backwards" to the management of documentation which was not originally planned.

of a huge set of data, in order to create a Multimedia GIS for the publication of an archaeological landscape and sites, containing all the typical geo-spatial information.

We should underline that the present project has three aims: 1. recovery and management oflarge and varied archives of old heterogeneous data 2. integration of these data in a system of constmction/ verification of hypothesis about the study of the material and of the environment 3. organisation of all the data collected in a GIS based system.

The first step of this work was to realize a prototype simulation model, able to give to the user a complete vision of the problematic of this region and to allow him to analyze, through the computer, the data from an excavation nowadays inappropriate to support a normal management. Thanks to the GIS-based system for the study of the excavation of Shahr-i Sokhta settlement and graveyard, the user can check the hypothesis proposed until now, but also explore some new hypotheses.

The use of satellite imagery begins after 1961 for military purposes. Since 1972, with the launch ofLandsat-1 by the USA, it becomes available for scientists and the public. The lsIAO projects have made extensive use of this imagery for the study of ancient civilizations in Middle Asia and the Arabian Peninsula.

According to the available data, the methodology consists of the following steps: checking and analysis of the existing archives; codification and integration of the different tables in a relational system for the management of databases (RDBMS); digitization in CAD systems of the excavated inhabited areas and Necropolis phase maps, using different layers for each room or for each single grave and geocoding these maps; realization ofGIS system for the analysis and the study of the excavated areas of the site; queries for the realization of phase, thematic and diachronic maps directly from the system (fig. 9) construction of a 3D model of the landscape and of the site of Shahr-i Sokhta (figsl0-11 ); visualization of the site with excavation areas geocoded (fig.12); texture mapping of the satellite photo on the model of the site ofShahr-i Sokhta (fig.13).

Our only serious possibility to understand the complex fluctuations of peoples and their socio-economic organization through time, since no occidental scientists can work in the area, rests only on the progress of satellite imagery for earth studies and related computer processing. The old surveyed data from the explorers of the l 9th century and the archaeologists of the 20 th will be projected on satellite images of highest resolution. For the Archaeological Map of the Helmand Delta, currently under preparation in Rome and Bologna, a wide collection of maps and satellite images has been gathered to define altitude variations and geo-morphological conditions at the highest level. Recently NASA has released the collection of the first militmy images from the CORONA Program, that are particularly impmiant because they date to the Sixties when Sistan was still very little developed (figs 5-6).

With this approach, it would be possible to arrive, at the end of the data elaboration processes, or the first "virtual publication" of the protohistoric site. Moreover, the development of such multidimensional GIS, will make possible a multimedia publication of the digitally elaborated archaeological infom1ation, in order to give the scientific public the possibility of exploring and understanding the fieldwork data and infom1ation for the entire site complex (i.e.: a final CD-ROM).

The idea of using GIS for the management, analysis and study of old and new data on Sistan will be directed towards reconstructing the landscape and environmental changes for the study of early civilizations in their territorial dimensions, beginning with Shahr-i Sokhta and the Early Bronze Age settlements. Central to the understanding of changes in the area is the erratic behaviour of the lake, whose fluctuations largely marked the wealth potentials of the region (figs7-8)

Condusion The methodofogy It is clear that we have presented here only the methodological

Since I 978 the Italian Expedition oflsIAO could not continue fieldwork activities in the Sistan region, and from that moment the digital processing phases of the huge and heterogeneous collection of data began. Until now, several tries of data recording, management, and visualization, such as a dedicated database, map digitization, and a Hypermedia system, developed in Too/Book, have been carried out.

approach to the problem of recovery of data and infonnation from an old excavation and survey. The first target we kept in mind was to find a system able to manage and record all the data and that allows us to study and to formulate some new interpretative hypotheses. Until now we have worked mainly on the collection and management of data, GIS as the engine in which all the information will be stored, in order to allow the researchers to produce quickly and easily some basic thematic maps and to check all the data .

The main aim of the project is to develop new methodological and technical approaches for the acquisition and the elaboration

The next will be the most interesting and'-"""''""'' as it will consist of the intra-site and of some analyses

87

of the working areas definition on the site. We started some analyses and at the moment the result is very satisfying, pushing us to formulate each day new intra-site analyses.

Laurenza, S. et alii: "Horse-Grave ofAvarian Age (7th. A.D.) in S.E. Hungary: A Virtual Flight Implementing Parallel Enviroment" in Barcelo, Forte Sanders (eds.)" Virtual Reality in Archaeology". BAR International Series 843, Oxford, 2000.

Our wish is to achieve a Multimedia GIS-based publication of the entire corpus of data relative to the study of the protohistoric site ofShahr-i Sokhta by next summer, which will allow us to present new and innovative hypotheses on the difficult problematic of this region.

Lock, G., Stancic, Z. 1995. Archaeology and Geographical Information Systems. Taylor & Francis , London

Acknowledgement

Lock, G. 2000. Beyond the Map : Archaeology and Spatial Technologies. In : Lock G. (ed) NATO Science Series, Series A : Life sciences - Vr_i/.321

The authors are grateful to Prof Maurizio Tosi who provided material, data and time to make this project possible.

Redman, C.R., Watson, P. J. 1971. Systematic, Intensive Surface Collections. American Antiquity. 35 (3): 319-325.

References

Renfrew, C., Zubrow, E. B. 1994. The ancient mind. Elements of cognitive archaeology, Cambridge, 1994.

Kirkby,A., Kirkby M. J. 1976. Geomorphic Processes and the Surface Survey of archaeological sites in Semi-arid Areas. In D. A. Davidson and M. L. Schackley (eds.), Geoarchaeology, London, pp. 229-253.

Tosi, M. 1985. Dalla "molle" Shahr-i Sokhta alla ''rigida" Moenjodaro, ovvero il consiglio disatteso. In Studi di Paletnologia in onore di Salvatore J\;f Puglisi. Roma, La Sapienza, pp. 277-289.

Johnson, I., Norton, M. 1996. Archaeological Applications of GIS. In : Proceedings of Colloquium II, UISPP XIIIth Congress, Forli, Italy

Rawlinson, H. R. 1873. Notes on Seistan. RSGjournal xliii, London, pp. 272-294. Salvatori, S. and Vidale, M. 1997. Shahr-i Sokhta 1975-1978: Central Quarters Excavations. In : Preliminary Report. Reports and Memoirs, Series Minor I, Roma: lsIAO

Laurenza, S. et alii. 1995. Un Sistema Ipermediale per navigare in un'antica citta' dell'Asia Mediorientale, Pubblication n.36, Aprile 1995, Collana IRSIP

Vidale, M. 1995. Viaggio intorno alla mia ciotola. Evoluzione tecnologica e comunicazione non verbale in una sequenza ceramica dell 'Eta del Bronzo. Annali dell 'Jstituto Universitario Orientate 55, 3, Suppl. 84 Napoli: Istituto Universitario Ori en tale

Laurenza, S. et alii. 1995 .. Surfing the Past: A tool for hypermedia visit of an early urban site in Middle Asia. In : 1st International Congress on : Science and Technology for Safeguard of Cultural Heritage in the Mediterranean Basin, Catania, pag.475

Vidale, M., Tosi M. 1996. The Development of Wheel Throwing at Shar-i Sokhta. Slow and Fast Revolutions towards Statehood. East and West 46, Rome: lsIAO, pp. 251-269

Laurenza, S. et alii. 1996. Sharh-i Sokhta, la Citta Bruciata. In M. Forte (ed.) Archeologia. Percorsi virtuali nelle civilta scomparse. pp.200-204

88

Figure 1. Helmand river viewfrom Landsat mosaic image

Figure 4. Areial photo of Eastern Residential Area of Sharh-I Sokhta

Figure 2. Potsherds on the top surface of Sharh-i Sokhta

Figure 5. lvlosaic of Corona satellite photos

Figure 3. Historic map of Sistan (!873)

Figure 6. Mosaic of Landsat satellite photos

89

Figure 7. Reconstruction of Sis tan base map with seasonal rivers net Figure 9. Queries on Burnt Building area of Sharlz-1 Sokhta

Figure] 0. GRID processed with 3D Analyst of Sistan basin

Figure 8. Overlay of vector ancient historical maps with modern reconstructed base map

Figure 11. 3D reconstruction of the site with investigated areas

90

Figure13. 3D view of the site with satellite photo draped on

Figure] 2. Geocoding of the excavated areas with satellite photo

91

Fr

Stratigraphic Unit to the mouse: a based system for the excavation historical complex. The case study Pompeii Sabatino Laurenza Universita degli Studi "La Sapienza" Roma, Italy E-mai 1:laurenza. s@tiscalinet .it

Cristiano Putzolu Universita degli Studi "La Sapienza" Roma, Italy E-mail: [email protected]

Abstract: The needs of archaeologists are the recording, the management and the elaboration of a huge quantity of data and it/formation linked to different monuments which put us in .fhmt of d(fficult questions. The systematic proceedings of research increasis the quantity of information to be recorded and managed. The archaeologists need a quick and correct elaboration of collected data, in order to understand and analyse the investigated contexts. In this.f1wnework, we developed our application, born under a MURST 1998 Project (Ministry o_fUniversity and of Scientific and Technological Research). The prC/ject, started in May-June 1999 at the University "La Sapienza" of Rome, under the Scientific Direction o_fPro_fAndrea Carandini,.focused on two case studies: Pompeii and Palatino, having as its main aim the realization of a methodology and o_fa G/S based system.for the recording and analysis o_fan archaeological excavation. Brie.fly, the system consists in the elaboration of a desktop G/S based system helpful.for the archaeologists during the excavation phases of recording and management and able to support SU.forms with as much information as possible. ft is mainly a system.for the management and.for the analysis of the stratigraphic sequence, both graphic and alpha-numeric. Moreover, a final 3-D visualisation o_fthe excavation is displayed, broken down into its elements, Stratigraphic Units, in order also to link the in.formation and the data coming out during the excavation to 3D SU. In this way, we hope to complete the usual graphical documentation (plans and prospects) with 3D display format, showing swfaces and volumes of excavated SU. Key words: Stratigraphic Units, Dbase, GIS. 3D modelling, TIN.

Introduction

the distribution of artefacts in sites to understand the organisation of daily life. On a Macro-scale, the position of sites in the landscape and relative to one-another informs us about social organisation and economic strategics.

Archaeology is the study of the human past through material remains and traces. Archaeological remains range from complex urban sites and monumental architecture through individual dwellings to small "portable" objects such as broken tools. Archaeology is, by its nature, fieldwork based and interdisciplinary. Its scope ranges from the sciences of C14 dating and chemical analyses, through environmental and behavioural c,1..,1,~11~-'-'", to historical and literary research and even art history. Along the way phenomenal, and ever increasing, quantities of data arc collected. Far from being concerned with the individual spectacular object, out of its cultural and physical context, archaeologists attempt to piece together evidence and extract information from a myriad small details: the interrelationships between objects, the soil from which they were unearthed, the position of archaeological sites in the landscape and their relationship with the environment. The work is painstaking. To archaeologists with bigger datasets, computers have become essentials for their work.

Archaeology uses stratigraphic methodologies, which allow us to manage wide and multistratificd excavation areas and to reconstruct scientifically histories generated by the analysis of monumental complexes. The needs we have are the recording, the management and the elaboration of a huge quantity of data and infmmation linked to different monuments which push us daily in front of difficult questions. The systematic proceedings of the research increases the quantity of infonnation to be recorded and managed. Archaeologists need a quick and correct elaboration of collected data, in order to understand and analyse the investigated contexts. Everyone who has directed or managed a large scale excavation project, knows how difficult it is to remain infonncd about daily progress by comparing the excavation diary and reports, plans, Stratigraphic Units (SU) cards, maps and sections, photos and electronic fonns.

Archaeology is a spatial discipline, and this is why today desktop mapping software is so important. On a Micro-scale, we use

For decades directors of excavation projects have tried to put

93

order in the wide data-set of different information, using database archives and CAD system.

an archive. with combo-boxes checked vocabularies that will help the user which must input data and avoiding redundancy o data and lexical problems. The same kind of analysis has been developed for the graphic documentation, analysing the existing stratum plans and beginning the vector acquisition of these.

In the last ten years, in Italy, we have witnessed the spread of GIS in order to record, manage and display interactively huge archaeological data-sets, with priority on spatial and geographical infmmation.

Once shared, managed and overlapped the infonnation on the GIS base, we could observe differences and coincidences of recorded data, and so decide when, where and how to make a more accurate and detailed level of lecture and interpretation. The main function of our CHSbased system will consist on the reconstruction of the existing relationship between archaeological layers and their spatial intra site distribution. In fact, the GIS as a representation of infonnation dropped by the physical reality, must stand out for its own capacities to collect all the objects existing in the physical world, making them suitable to the mixes that really occur in the reality.

Between 1985 and 2000, under the scientific direction of Prof A. Carandini, research on the northern edges of Palatino and on Pompeii Insulae Vlll,2 and Vil, 9-11 carried out. Using funding from the Italian government (Ministero per l'Universita e la Ricerca Scientifica e Tecnologica, MURST) from 1999 we started planning for the realisation of a GISbased system able to manage the entire archaeological data-set, making self operating information recording and elaboration, in order to facilitate the reconstruction of phase to phase image sets for historical interpretation.

During the vector analyses the archaeological record is known to the operator since from the first phase of the project, in such a case the computing application asks, by the beginning phase, the accurate definition of an operative route able to normalize on distinctive levels the different existing information and the final goals to reach: the reconstruction of a topographic base, the overlapping of the Stratigraphic Units in a unique file, the recording of data linked then to the spatial information.

Excavation data, archives and the GIS engine The first target of our project was to have at our disposal, at the end of each fieldwork season, all the archives updated for all the data recorded in the field (graphic and texts) from which to extract automatically thematic diachronic plans (phase plans).

Beginning with such methodological remarks, we planned a research route marked by the following working steps : Coding of the infonnation and structuring of the field excavation data by realising of one database; Elaboration of thesaurus lexicon and of scripts of control routines in order to guide the data-entry phase; Digitise and transformation in vector of the cadastral and/or photogrametric plans and of all the stratum plans, each with different layers and SU; Geo-coding and assignment of a numeric identification key to the graphic objects which define the single SU; Integration of graphic and numeric/textual data, with spatial and not spatial attributes, by using a vector GIS as engine; Realisation of a graphical user-friendly interface (GUI) able to simplify the operations of recording and visualisation of the data in the database.

Considering that usual systems of recording archaeological excavations allow the recording of infommtion and limited graphic elaboration, and as we want to elaborate the stratigraphic sequence for separate phases creating thematic images, as the result of the combining of graphic, spatial and written information, we decided to choose to adapt the capacity of the GIS system to our needs. The system realised is based on a simple structure, mirroring the procedure of scientific elaboration of data collected in the field (Fig. 1). Starting from this point of view an option not to be set aside is that the system realised, based on interand common file formats, which allows easy exchange with colleagues, giving us a warranty of an opening for future solutions and an insurance for the data. Looking at the future, we decided to use from the beginning the . mdb format for data-entry, and to use Visual Basic release 6.0 for all the scripts, programming operations and also for the realisation of the database forms. It could seem too pretentious to say that almost all software houses, after a period of transition, are waiting for a more fixed and stable release of the Win '98 operating to make some~ .. ~.. ,"~"in their product, at least for the script language. It as everyone now knows, also the case for ESRI, which in the new release of ArcView 8.1 and of Arcinfo 8.1 changed the script language from Avenue to Visual Basic, and also with OLE/COM languages.

So, the work has been necessarily divided in different the first one is the planning of the archives, and it took away a long time for the analysis, also to evaluate the problems connected to the data coding and to define the logical and physical structure of the system (fig. 2). Considering also that infonnation must be organised in a way to satisfy different users at different layers, it seems to us obvious to choose a relational architecture, mirroring the elaboration of more archives connected between them by a primary key, here identified with the SU numbers. First archive contains all the data of SU, following the form of Italian Ministry of Cultural Heritage; the second one is assigned to the recording of graphic documentation (photos and drawings), by keeping always the relational link with the primary key, the SU number; the third one, contains each object found in the context and it is also linked to the primary key. In order

Considering the difficulties represented by the quantity and the diversity of data, an important step was the analysis of the existing data, both on paper (SU cards) and graphic/ topographic data (plans, maps, diagrams). So we decided to integrate the paper cards of the different samples chosen, by realising a specific thesaurus for each category of data, in order to realise

94

to make easier the data-entry for the students, we realised a database system which manages in a hierarchical mode the different tables, and that using graphic user friendly interfaces (GUI), drives the user during the choice of alternative values remotely controlled between them during the programming phase (figs 3-5).

of borderlines and of internal surface points and the photos of each SU.

The next steps of our project were represented by the vector transformation of the 1:20 excavation plans and by the realisation of the final GIS engine.

At this point, our problem was to optimise the survey proceedings and times; we didn't need a long and complex recording methodology.

The vector transformation of all the excavation plans was necessary to allow an exact collocation of the excavated strata and of some finds, and for the entire visualization of the existing graphic information, without compromising the overall readability of the data. So we proceed to transform in vector the different papery overlays, or better defined as stratum plans, beginning in this way the construction of the cartographic basic system, with a basic level of plans of stratum at 1:20 scale.

3D SSUU modelling

If we consider the SU volumetric value as the space between its surface and the surfaces of the SU covered by it, our system to record the complete three dimension of SU could be limited to the survey of the surface of each SU. We decided in this way to document the excavation of the room using an ETS, for the recording of the contours and of the surfaces of each SU. The SU were surveyed with an average of 100 points per square meter. Other interesting problem was the borderline: in fact, the surface of SU, which we will call upper surface from this moment, touches the underlying SU surfaces, called SU bottom, in a portion of space limited to it. At this moment we realised that the orthorectification of the digital images of SU was necessary; once orthorectified, thanks to the four control points surveyed by the ETS, the photo was georeferred and linkable to the topographic data of SU, following the same co-ordinate system. We decided to use an Arc View extension, 3D Analyst, to build the TIN of upper surface and of bottom of each SU; as those TIN overlap on the borderline, we could see the SU as an unique solid. On the upper surface TIN we overlaid as texture the georeferred and orthorectified photo. In this way we obtained a jpeg photo, including the values of the elevations and two different TIN with inside the volumetric value.

Established that the GIS engine works well for the management of 2D SSUU, we decided to try to accomplish 3D models of some SSUU in room VII of" Dom us della Pescatrice" in order to arrive to calculate the volume of each SU and to directly reconstruct, by the computer, the excavation phases.

Finally, the volumetric count was made by calculating the space between a before surface and an after surface, using the cut.fill function of ARCVIEW, where with before surface we mean the until now called upper surface and with after surface we mean the bottom.

Thanks to our methodological approach, that appears to be soon he right one and also thanks to the module 3D Analyst release 1.0 of the ESRI Arc View release 3.2a, it has been possible to follow the excavation phases directly on the field and than to reconstruct the excavation on a Desktop Personal Computer.

We must underline also that importing in a CAD the orthorectified and georeferred photo and points we could anyway produce a traditional plan, avoiding errors usual during the manual survey.

The post-processing phase is based on two data fonnats: the .DXF coming out from the total station and the .tiff digital images.

After this step, the GIS engine realisation starts, by linking together the data-entry and the graphic overlays of each SU, giving us already the possibility to extract phase plans directly by writing a simple SQL query (figs 6-7). At this point, once having done some joining and linking operations directly in the ESRI Arc View release 3.2a GIS, we try to extract some charts and analyses on the SU data, for example, the diagram of pottery by different periods, or queries about SSUU with some special finds (figs 8-10).

Because of the absence of findings and remains in the stratigraphic deposit, we couldn't make distribution analyses in this room; in fact surveying the location position of the remains in the SU deposit, we could easily visualise and analyse the primary contexts.

The main goal of the last fieldwork campaign in Pompeii was therefore to test a new standard of documentation, consisting in a research tool. Hardly the traditional documentation of an archaeological excavation recorded all the three dimensions of the space, in the best cases restricted to 2,5 D! The few elevations marked on the excavation plans and sections are in fact never satisfactory, because they can not represent the depth of the entire SU surface. Starting from this need, we elaborated a qualified strategy for 3D fieldwork data acquisition. We used an electronic total station and a Nikon digital photo camera for the data recording. considering that the documentation strategy included the survey of control points,

At the end, we want underline that often plans and sections not enough to represent the precise and entire aspect of investigated deposit, and they add only the physical dimensional data of the SU to the Harris matrix and to the recorded cards.

are the biSU

We think that this recording methodology could provide the archaeologists a set of 3D data, including the advanced possibilities of volumetric counting (figs 11-14).

95

Some final considerations

and future approaches

option.

In this paper we follow almost the same pattern of some our previous papers, about the Pompeii Excavation GIS, where we presented our system as a work in progress, with some results already obtained but capable of many implementations.

If one thinks of this system as a two-faced Janus, where each object is at the same time a database record and a graphic entity, one can easily understand that the major problems come from the graphic aspect.

In the present paper we tried briefly to summarise some of the results: The GIS database engine provides a solid archive for the thousands of US cards and the other papery documentation. Once in the database, it is easy to query (and to obtain answers if the queries are appropriate) this previously not nimble Leviathan. The digitising and geo-coding of the numerous overlays is not only a good procedure of archiving but allows the user to better visualise the relationships between the surveyed structures. From its beginning (i.e. without any GIS based operation but simply after the data entry of US cards) the system can fill automatically the Ministerial form for US documentation. After a simple query the system can provide the user with phase or period maps. 3D reconstructed SU with draped images give the possibility to calculate volumetric values and to visualise in a more realistic way the aspect of the excavated ( i.e. destroyed) deposit. Even if still in progress our GIS already allows simple and crossed SQL queries both from the tables of the database and from the graphic representations of the structures.

The planning and creation of the new database for Wall SU data-entry was made easier by the experiences gained with the previous SU one. This doesn't mean that we duplicate the database we already have and use it for Wall SU but that now we know exactly how to plot the diagram of the relations between the tables, how to prepare the appropriate user-friendly layouts, how to control input data errors, just because we made some errors and wasted time creating the SU data-entry! At the beginning, we hoped the same for the graphic aspect : utinam id sit. quod spero ! (Terentius). But we realised immediately that Arc View 3D Analyst 1.0 doesn't allow in any x-z or y-z plane the same data processing it uses for x-y plane. In fact, TINs creating method builds up a surface joining each point with the ones contiguous: stricter is the net of points, more accurate is the surface obtained. The other factor involved in TINs creating is the boundary, a line that circumscribes the area where the new surface will be created. In the case of SU, the boundary lays in the x-y plane and the software have no problems in TIN s creating; but working with Wall SU, the points and the boundary of the surface belong to a x-z I y-z plane polygon.

But we would like to illustrate briefly the main implementations we are working for: l. We want to extend the 3D modelling to the walls and mainly to the study of the structures, as for the reconstructed SU, making possible queries directly from the graphic interface of our GIS. 2. We are trying to realise automatically the Harris Matrix and to keep it visible and editable directly from the GIS engine.

3D Analyst manages that data as if they lays in the x-y plane; it seems to us that the software processes only surface points with x and y coordinates that fall inside the boundary line projection on the x-y plane. In fact, in our tests, we can see that there are some points involved in the process of TIN creating and others (often close to the fmmer ones) totally ignored.

The problems to solve for those two points are still many, even though we have already reached some good results from the research point of view.

3D managing of the Wall Stratigraphic Units

It seems at the same time that even the neighbourhood between surface points is calculated in the x-y plane (i.e. ignoring the elevation of each point) so that often the triangles of the surface are created from points that in reality are not so close! (figs 15-17)

1

Now it's important to keep in mind what we want from our system: we don't want to bring out one of the thousands of 3D Virtual Trip of Pompeii and we don't want to digitally rebuild our domus case study; we just want to manage the huge mass of data coming from the excavation and gain new knowledge analysing them. This is not a secondary aspect: if we already have Wall SU data from the database and, at the moment4, our GIS software architecture doesn't allow a photo realistic 3D reconstruction of the walls, we can consider satisfied with the simple Wall SU

Due to the scarcity of resources available this year for the entire project, we can here present only the preliminary stages of our work that, in our intentions, will push the quality of USM data managing at the same level reached for SU. It is important here to remember that our GIS is entirely developed in ESRI Arc View 2 because we think that the winning choice for this kind of systems could be the complete intercommunication between each module 3 ; that being stated 3 D Analyst 1.0 extension is, at the moment, the only available

96

extrusion.

period.

This solution was tested and provides a valid graphic interface for Wall SU database in the GIS 3D scene.

Keeping in mind all these rules and laws, we are trying from this year to realise an application able to reproduce directly on computer the Harris matrix: as all our programming language of our database is entirely made in Visual Basic release 6.0, we prefer to continue to use it also for the reproduction of SU diagram. We decided to realise an OCX applet, able to be connected directly to different database fmmats, i.e. dbf, mdb, etc., and that mainly will be able to read directly the data from the interesting fields and to reproduce the diagram.

With the extrusion of our bi-dimensional plan of Room VII we were able to extend our GIS analyses to the structures in elevation, making queries directly on 3D structures as on the 2D views. Anyway, we must recognise that in our case of study, the stratigraphic sequence of the walls is quite easy, with each Wall SU one beside the others, and that in case of a more complex sequence (with Wall SU one above the other) could be hardly visualised by a simple extrusion process (fig. 18).

Until now we have developed the OCX applet that works well for the main stratigraphic rules; we are now working in order to update and to extend the package, making possible automatically the calculation of all the redundancies (fig. 19). Moreover, as already told before in the text, with the new version of Arc View release 8.1, we would like to make a porting of all the programming scripts made in Visual Basic directly in the GIS system, and to allow the user to have also the possibility to open the matrix during the vie of the excavation and to edit it directly by inside the system. It is hard work but we can say that is more and more possible, especially by the fact that with new Geodatabase structure of Arc View, we could establish some fixed relationships between graphical objects, in our case the SU.

The Harris Matrix As usually said, the work of the archaeologist seems to be quite a strange activity; in fact, during a first step of his work on field he must excavate and so to remove the stratigraphy, that, in a second phase of his work he will be obliged to rebuild during the elaboration and the interpretation of data. Such a model has a its own graphic representation, the so called" Harris Matrix" or also the "stratigraphic diagram". It is necessary to resume all the existing relations, that in case of monumental complexes excavations can be also thousands, between the SU or the Actions in them identified. The layers are the most little unit of an archaeological identification and they have besides their spatial dimensions also a time (chronological) dimension. As a result Harris developed the so called" Harris Matrix" able to describe and to represent graphically the time relationships between different layers. We could imagine that the layers in a Harris Matrix are presented as a rectangle containing the layer name (or number). The relations draft as lines and the position of linked rectangles describe the type of relation. Such a graphic representation corresponds well to the common image of archaeological excavations, where it is usually expected that the most recent layer is above to the oldest one. As there are several exceptions, the relations must always be confinned by some other observations.

Anyway, we hope to finish this application for the next CAA 2002 in order to present the full operative system applied also on other excavations ( i.e. a protohistoric site).

Conclusion At the beginning of the HI millennium we can say that the Information Technologies are inside all the aspects of our work. The need to show to the others and to check ourselves directly in field the data and the information collected during an excavation is nowadays possible, thanks to the IT. Archaeologists are using the modern computer systems in the field since from the beginning of 1980, receiving a grateful help for the management and the recording of the information. Since from the '90s the diffusion on wide scale of the GIS systems give the possibility to manage not only the textual infonnation but also the graphic aspects of it, linked directly each others and mainly to push the intra-site analyses during the same excavation phases.

From the mathematic point of view, the relation " After than" could be defined as a set partially ordered. So, we could easily define some additional rules for the relations: Relations are thoughtless, asymmetric, transitive and anti-cyclic. For the relation "Contemporaneous to" we have different rules: These relations are reflexives, symmetric, transitive, and they are called "relations of equivalence".

The use of GIS for excavation is today a reality, also because the specialization of some archaeologists in GIS use is increasing.

\Ve don't want here to resume the concepts of Harris Matrix but just we would like to underline that anyone which could realise a software or an application for the matrix diagram must be very careful to observe these basic rules. In order to increase the readability of the Harris matrix it will be useful to clean it from all the so called "redundant relationships". These are relations already formed by other relations. The stratigraphic diagrams or matrix allow therefore to represent in a very schematic way on a bi-dimensional plain the 3D reality of the stratification of an archaeological deposit, and the single SU wiU contribute to define the phases of an historical

Starting from this consideration we can affirm that during the realization of an excavation GIS is better to think to the archaeologist's questions than to the problems of IT experts, making to himself the most usual questions : What is an excavation GISfor? How wi!l it be made? Which benefits could I get from it? Could it be a tool usedfor analytical queries and for the making of historical and interpretative models? When we started, our main aim was to create a system useful

97

for the management and for the analysis of the excavation archaeological record; after two years, we could say that not only have we accomplished the system but also that today at the University Laboratory there is a team able to manage computers and to use it for its own needs.

archaeological research and for didactical purposes.

Acknowledgements Special thanks to Prof P. Carafa for the Scientific coordination of the project. The authors would like to thanks all the students who participated to the data-entry and to the vector transformation of the SU plans.

Our work is based on three main and important assumptions : The computer is only a tool in the hands of the archaeologist; We don't have to approach the computer assisted systems with fear of them; We must be able to programming scripts, in order to create specific applications.

End notes 1 In Italy abbreviated as USM (Unita Stratigrafica Muraria, i.e. Wall Stratigraphical Unit). 2 We developed the database with MS Access and Visual Basic and we digitised the overlays with AutoCAD R. 14 in order to provide the system with tabular and graphical data but, once created the system, we tried to exploit Arc View resources as, for example, the 3D UUSS reconstruction realised in Arc View 3DAnalyst. 3 The latest release of ESRI Arc View 8 .1 seems to confirm this philosophy with implemented graphic and database (geodatabase) capabilities and the choice of Visual Basic instead of Avenue as internal programming language. 4 We are waiting for the new 3 D extension ofESRIArcGIS 8.1 that has been distributed in Europe only since last month.

Our final system is not the ultimate one, but at the beginning of the new millennium we must consider that the winning philosophy of computer systems is represented by personalised solutions. We can say that our system is functional and useful for the management of different excavations, even if it would be not the absolute one. The real problem today in Italian Universities is to decide what could be the computer knowledge degree of archaeologists. Archaeologists must know directly the recording and management processes; the computers used today give us the possibility to do it. However, it requires the diffusion in the Italian University Departments of basic computer know-how and the realisation of a theoretical and practical school. Archaeology could not survive without the use of computers, but would risk being kept out of new communication systems, which always demand complete and transparent documentation, high and fast transfer protocols, and different reading keys.

References Allen, K. M. S., Green, S. W., Zubrow, E. B. 1990. Interpreting Space : G!S and Archaeology. London: Taylor & Francis.

The specialists and also common people must have the possibility to access to publication and visualisation of the researches to which they are interested. If we are not able to adapt ourselves to modern communication systems and to transfer our data to different subjects, we will be out of the game and we will work only for ourselves. Archaeology is perfect for digital technology; it finds in this branch several spectacular applications, which will help us to follow modern dynamics and language based essentially on the images. The problem of image language is to keep the distance between ''.false true" and "truefalse". No reconstruction will give us the exact image of the past. This gap between virtual reality and hypotheses on the past is a constant in a inexact science as archaeology, with which the researchers use to play. The final image must increase the value of the scientific reflection and it must exist as a step of the archaeological research process.

Alvey, B. A. P. 1993. Interpreting Archaeology with Hindsight: the use of three dimensions in graphic recording and site analysis. In E. C. Harris/m. R. Brown III/G. J. Brown (eds), Practices q/ Archaeological Stratigraphy, 218-228. London: Academic Press. Andersen-Madsen. 1996. IDEA - the integrated database for excavation analysis. In Interfacing the Past, CAA 95, Annalecta Prehistorica Leidensia 28, pp. 3-14. Barcelo J. A., Forte M., Sanders D. A. (eds). 2000. Virtual Reality in Archaeology. BAR International Series 843, Oxford. Beex, W. 1995. From excavation drawing to archaeological playground: CAD applications for excavations. In: J. Wilcock/ K. Lockyear (eds.) Computer Applications and Quantitative Methods in Archaeology 1993, 101-108, BAR International Series 598, Oxford : Tempus Reparatum.

IT therefore represents one of the new tools for archaeology and it must be used, as such, with care and attention, exactly as the trowel.

Biswell Cropper Evans J., Gaffney V., Leach P. 1995. GIS and excavation: a cautionary tale from Shepton Mallet, Somerset, England. In G. Lock, Z. Stancic (edds.) Archaeology and Geographical Information Systems: a European perspective. London, pp. 21I-229.

The ambitious purpose of this project, if such exists, is not to open new frontiers for the recording and the management of archaeological record. but to suggest some guidelines and methodological approaches to the excavation data management. At the end, we hope that the development and future extensions of this system could be a suggest for the institutions to begin to consider seriously those new and powerful tools for the

Carafa P. 1997. What was Pompeii before 200 be? Excavation in the House of "Joseph II" and of House of·'nozze di Ercole". In

98

: S. Bon/R. Jones (eds.) Sequence and Space in Pompeii, Oxford, pp. 13-31.

Paradis.A. R., Becker, R. C. 1990. Interactive volume modellinga new product for 3-D modelling. Geobyte 5 (1), 42-44.

Carandini A. 1995. Storie della terra, Torino, Jrded.

Pigat, Hazelton, B. 1992. The fundamental of a topological model for a 4-D GIS. Procs. 5 th International Symposium on Spatial Data Handing 2, 580-590.

Hardy, R. L. 1988. Concepts and results of mapping in three dimensional space, Technical Papers ACSM-ASPRS 1988 Annual Conference,2. Baltimore, MD, 106-115.

Reilly, P. 1992. Three dimensional modelling and primary archaeological data. In : P. Reilly/ S. Rahtz (eds), Archaeology and the Information Age. A Global Perspective, One World Archaeology 21, 147-173. London: Routledge.

Harris, T. 1988. Digital Terrain Modelling and threedimensional surface graphics for landscape and site analysis in archaeology. In: C. L. N. Ruggles/S. P. Rahtz (eds) Computer

Application and Quantitative Methods in Archaeology 1987,

Rick, J. W. 1998. Total Stations in Archaeology. SAABullettin, 14( 4) : Interface-total station.

161-170, BAR International Series 393, Oxford: Tempus Reparatum.

Tucker, G. Lancaster, S. T., Gasparini N. M., Bras R. L., and Rybarczyk, S. M. 1999. An object-oriented framework for distributed hydrologic and geomorphic modeling using triangulated irregular networks. Computers and Geosciences in press.

Harris, T. M. and Lock, G. R. 1995. Multi-dimensional CHS: exploratory approaches to spatial and temporal relationships within archaeological stratigraphy. In : Inte1facing the Past, Annalecta Praheistorica Leidensia 28. CAA 95,

Tucker, G. E., Lancaster, S. Gasparini N. M., Bras R. L., Zeidler, J., Johnson, W., and Isaacson, J. 2000. Modelling the 3D stratigraphic context of Prehistoric sites : a new approach using process-based computer simulation. Paper presented at the 2000 SAA Conference. Philadelphia (USA).

Hinshelwood-Dalla Bona. 1994. GIS and lntrasiteAnalysis: An Example from Northwestern Ontario. Archaeological Computing Newsletter 40, pp 12-20. Jones, T. A., Leonard, J. F. 1990. Why 3-D modelling? Geobyte 5 (1), 25-26.

Wunsch, G. , Arasa , E. and Perez, M. 1995. Dissecting the palimpest : an easy computer graghic approach to the stratigraphic sequence of Tunel VII site (Tierra del fuego) Argentina. SAA

Laurenza, S. and Putzolu, C. 2001. Sistema lnformativo Dinamico per la gestione e l'analisi dei dati archeologici : l'esempio di Pompei. In : Usi e Consumi dell' informazione geografica 2001, 209-217, I Quaderni di MondoGIS, Atti della Terza Conferenza di MondoGIS, Roma.

Youngman, C. 1989. Spatial data structures for modelling subsurface features. In : Raper J. F. (ed.) Three dimensional applications in Geographic Information Systems, 129-138. London: Taylor & Francis.

Maggiolo Schettini et al. 1995. Computation and representation of stratigraphic sequences in a system for archaeological data, Archeologia e Calcolatori 6, pp. 173-187.

99

Figures

Figures nt:1.,0 W(I IU( • -O.\C:AV/l.nuf'0.1 p>0.1 p>0.1 p>0.1 p>0.1 r = 0.397 r = 0.142 r = -0.087 r = -0.060 r = -0.255 r = 0.395 South n = 23 n = 23 n = 23 n = 23 n = 10 America n = 23 0.1>p>0.05 p>0.1 p>0.1 p>0.1 p>0.1 p>0.1 r = -0.390 r = -0.390 r = -0.017 r = 0.000 r = -0.387 Insufficient Asia n = 22 n = 22 n = 22 n = 24 n = 24 data. 0.1>p>0.05 0.1>p>0.05 p>0.1 p>0.1 0.1>p>0.05 Table 1. Results of regression of innovation, tool use and social learning frequencies, and measures of brain size, on total geographical range size, using phylogenetically independent contrasts.

Innovation Tool use Social Absolute Relative Neocortex frequency frequency learning brain size brain size ratio r = 0.141 r = -0.161 r = 0.233 r = 0.059 r = 0.063 r = 0.030 Africa n = 24 n = 24 n = 24 n = 33 n = 33 n = 12 p>0.1 p>0.1 p>0.1 p>0.1 p>0.1 p>0.1 r = 0.397 r = 0.142 r = -0.087 r = -0.060 r = -0.255 r = 0.395 South n = 23 n = 23 n = 23 n = 23 n = 10 America n = 23 0.1>p>0.05 p>0.1 p>0.1 p>0.1 p>0.1 p>0.1 r = -0.390 r = -0.390 r = -0.017 r = 0.000 r = -0.387 Insufficient Asia n = 22 n = 22 n = 22 n = 24 n = 24 data. 0.1>p>0.05 0.1>p>0.05 p>0.1 p>0.1 0.1>p>0.05 Table 1. Results of regression of innovation, tool use and social learning frequencies, and measures of brain size, on total geographical range size, using phylogenetically independent contrasts.

Dependent variable Spatial variability

Seasonal variability

Interannual variability

Location Africa (n = 24)

Innovation frequency r = 0.505 p0.1

r = -0.124 p>0.1

r = -0.210 p>0.1

Africa

r = 0.396 p = 0.05

r = 0.305 p>0.1

r = 0.329 p>0.1

South America

r = -0.152 p>0.1

r = 0.022 p>0.1

r = -0.147 p>0.1

Africa

r = 0.376 0.1>p>0.05

r = 0.447 p0.1

South America

r = -0.312 p>0.1

r = -0.123 p>0.1

r = -0.147 p>0.1

Dependent variable Spatial variability

Seasonal variability

Interannual variability

Location Africa (n = 24)

Innovation frequency r = 0.505 p0.1

r = -0.124 p>0.1

r = -0.210 p>0.1

Africa

r = 0.396 p = 0.05

r = 0.305 p>0.1

r = 0.329 p>0.1

South America

r = -0.152 p>0.1

r = 0.022 p>0.1

r = -0.147 p>0.1

Africa

r = 0.376 0.1>p>0.05

r = 0.447 p0.1

South America

r = -0.312 p>0.1

r = -0.123 p>0.1

r = -0.147 p>0.1

Table 2. Results of regression of innovation, tool use and social learning frequencies on measures of climatic variability, using phylogenetically independent contrasts.

Table 2. Results of regression of innovation, tool use and social learning frequencies on measures of climatic variability, using phylogenetically independent contrasts.

Figures

Figures

Number of species

- 4-7 □ 1-3

Figure 1. Primate species richness in Africa, from Wolfheim (1983). 112

-

8-11 12-15

-

16-19

Figure 1. Primate species richness in Africa, from Wolfheim (1983). 112

The Creation and Potential Applications of a 3-Dimensional the Early Hominin Site of Swartkrans, South Africa Joseph D. Nigro Center for Advanced Spatial Technologies / Department of Anthropology University of Arkansas Ozark Hall, Room 12 Fayetteville, AR 72701, USA Phone: + l-501-582-0790 - Fax: + l-501-575-5218 - E-mail: [email protected]

W. Fredrick Limp Center for Advanced Spatial Technologies/ Department of Anthropology University of Arkansas Ozark Hall, Room 12 Fayetteville, AR 72701, USA Phone: +1-501-575-7909 - Fax: +l-501-575-5218 - E-mail: [email protected]

Kenneth K. Kvamme Department of Anthropology University of Arkansas Old Main 330 Fayetteville, AR 72701, USA Phone: +l-501-575-4130 - Fax: +1-501-575-6595 - E-mail: [email protected]

Darryl J. de Ruiter Palaeoanthropology Unit for Research and Exploration Department of Palaeontology Bernard Price Institute for Palaeontology Private Bag 3 University of the Witwatersrand 2050 Johannesburg, South Africa Phone: +27-11-717-6668 - Fax: +27-11-339-7202 - E-mail: [email protected]

Lee R. Berger Palaeoanthropology Unit for Research and Exploration Department of Palaeontology Bernard Price Institute for Palaeontology Private Bag 3 University of the Witwatersrand 2050 Johannesburg, South Africa Phone: +27-11-717-6668 - Fax: +27-11-339-7202- E-mail: [email protected]

Abstract: The accumulation offossil remains, bone tools, and stone tools at the site Swartkrans has been attributed to a variety qf agents including hominin activity, carnivore activity, alluvial deposition, and gravitation. In.fact, the accumulations discovered at Swartkrans and other Plio-Pleistocene cave sites in South Africa have probably re.rnltedfi'om a combination of these.factors. In order to explore further the taphonomic nature of this deposit a Geographic Information System (GIS) was constructed incorporating all of the information produced.from the Swartkrans excavations and a recent survey of the site. The final product represents a digital archive of Swartkrans data that allows the user to simultaneously visualize and analyze fossil, artifact, and geological data together ivithin their spatial contexts allowing a more comprehensive investigation qf the site postexcavation. The mapping and 3D reconstruction of sites such as Swartkrans present challenges when using traditional Ci!S approaches. These issues are addressed, as well as the potential applications of the system. Key words: Geographic Information Systems, GJS, palaeoanthroplogy, hominin, hominid, taphonomy,fossil assemblages, intra site analysis, Plio-Pleistocme caves, Swartkrans

113

Introduction

Provenience data of the fossils and artifacts

Travertine caves in the Sterkfontein Valley of South Africa's Gauteng province have yielded a rich array of Plio-Pleistocene faunal material. The site of Swartkrans, for example, has produced the single largest sample of fossils attributable to the genus Paranthropus in addition to numerous fossils belonging to the genus Homo as well as a variety of bone and stone tools. To date, approximately 350,000 pieces of fossil bone representative of both human and non-human species (Brain 1993:3 ), along with almost 900 stone tools and 70 bone tools, have been discovered at this site (Brain/Shipman 1993:214).

Palaeoanthropological exploration commenced at Swartkrans in 1948, but it wasn't until 1979 when systematic excavations were first conducted at the site. Many important specimens came to light during the first three decades of work there, including the well-known cranium of an adult Paranthropus robustus (SK 48). Unfortunately, all these early discoveries arc lacking the vital provenience information needed to accurately reposition them in three-dimensional space (Brain 1981: 221). This is an impmiant fact because it hinders us from producing a total reconstruction of the geological and fossil content of Swartkrans.

The decisive work of C.K. Brain, conducted between 1965 and 1986, has provided many important clues concerning the accumulation of the fauna at Swartkrans. Based on carnivore markings on the bones, as well as geological fommtion processes, Brain suggests that the fossil accumulations of Swartkrans might be best attributed to carnivore activity (Brain 1968, 1969, 1970, 1974, 1981, 1993). Recent studies have yielded an array of possible hominid-derived tools and have, by implication, raised questions concerning the formation processes responsible for their deposition (Clark 1993; Brain/Shipman 1993; Backwell/d'Errico 2001) drawn from the spatial relations of the remains at the site. In order to test the hypothesis that a variety of agents were responsible for the accumulations at Swartkrans and to extract and isolate these complex agents of accumulation, a Geographic Infonnation (GIS) was constructed.

In 1979, this problem was remedied with the establishment of a datum, in the fonn of a pcnnancnt metal grid, over the Outer Cave area (fig. I). Although the main bars of the grid were constructed of rigid metal, the smaller units were made of wire and spaced every meter. Over areas representing Members (strata) I and 2 each grid square was divided into 50 cm 2 quadrants while subdivided even further into 25 cm 2 subquadrants over Member 3 due to the density of burnt bone found in twenty consecutive spits throughout a 6-m vertical profile (Brain 1993: 4). The excavation was mainly carried out in 10cm spits in all areas of the Outer Cave, but extending as deep as lm in certain places.

Mapping the existing site

In 1999 the Palaeoanthropology Unit for Research and Exploration at the University of the Witwatersrand in conjunction with the University of Arkansas' department of Anthropology and it's Center for Advanced Spatial Technologies assigned the first author to the task of developing a GIS for Swartkrans. The project arose as an exercise in applying recent technological developments to a site that was excavated at a time when GIS was either in its infancy in regards to archaeological applications or not an option at all. Swartkrans was used as the pilot site to demonstrate how this technology is capable of deriving new or enhanced information from w,.,,, .... " data and facilitating current methods of analyses. It was chosen because of its relatively small size due and the existence of meticulously recorded fossil, artifact, and geological data collected during C.K. Brain's excavations.

The first stage of the project involved mapping Swartkrans as it exists today. C.K. Brain was asked to visit the site to assess if there had been any major shifts in the positions of the remaining geological structures since his first removal of excavated material, material that may have supported these structures previously. Fortunately, no major changes had taken place and so the site as it exists today reflects the site upon initial excavation in 1979 minus the excavated matrix (Brain 1999). This was useful information since the purpose of the survey was to incorporate all of the existing geology of the site into the GIS in an attempt to increase the recorded provenience of the lmown fossils and artifacts. For example, if a fossil was recorded to have come from the NW quadrant of the grid-based position E4S7, the accuracy of the provenience would be 50 cm 2 x 10 cm. By mapping existing structures, the accuracy of the provenience information could be increased if uncxcavated material still remains in E4S7 thus decreasing the volume where the fossil was originally found to less than 50 cm 2 x 10 cm (fig.2). These structures were sometimes recorded in the excavation notes, but mapping the site filled the data gaps where representative diagrams are non-existent. In addition to increasing the provenience data of the finds, the mapping of Swartkrans provides both a three-dimensional matrix into which the excavated data can be repositioned and a complimentary source of geological infonnation derived from the attributed survey points.

The specific goals of the project were: 1) to create a system that can assist in further understanding the stratigraphy present at Swartkrans and other cave sites; 2) to create a system that can be used to further explore the relationships between the fossil and artifact distributions of these sites; 3) to realize the implications and constraints of the technology involved in creating such a 4) to construct a 3D model of Swartkrans in order to improve the grid-based provenience information of the finds recorded previously; and 5) to reconstruct C.K. Brain's excavation. The Swartkrans GIS was demonstrated to be very beneficial in addressing all of the goals listed above and has resulted in the creation of a digital archive encompassing all available site information. This accomplishment wiU allow researchers to virtually re-excavate the site an infinite number of times. a task that would otherwise prove impossible.

The survey was caffied out using a Wild TlOHJTMtheodolite with a Leica Distomat TM Laser electronic distance meter attachment (fig.3 ). A total of 14,885 survey points were collected representing the Inner, Outer, and Lower cave systems.

114

The Outer Cave, now a sinkhole due to collapsed dolomite roof blocks, was the only area recorded with the intention of creating a 3D reconstruction, given available provenience infonnation. The base perimeters of the Inner and Lower caves, however, were recorded to determine their spatial orientations within the cave network.

record with a laser than with a camera. In addition, a method that can be used in dark, subterranean networks needed to be devised in order to extend this work to other cave sites.

Constructing the Inner and Lower perimeters and the excavation grid

The survey was based on C.K. Brain's grid system to facilitate the incorporation of the data derived from his excavation notes within the 3D model. As a result, the origin (represented by the center of the grid structure) and the north line (the metal bar leading from the origin to the ''Hanging Remnant", a landmark feature) remained the same as those used during the excavations (fig.4 ). One obstacle encountered at this point was the problem of dealing with the wire section of the grid where the wires were either sagging or gone completely. This made it impossible to map these grid sections and, as a result, each digitized excavation diagram was unable to be reprojected individually to its original location. Fortunately, the metal bars constructing the main axes of the grid were still in place and so all of the excavated squares were reprojected as one unit based on these axes.

Cave base

The survey points representing the base perimeters of the Inner and Lower caves, as well as Brain's excavation grid, were imported into AutoCAD MAP R3. Each consecutive point was connected by a line using the 3D-polyline command, thus producing a three-dimensional vector representation of each feature. Upon completion these layers were saved as data interchange files (.dxf) and imported into Arc View 3.2 using the CAD Reader extension. It was essential that these layers were saved in .dxf format, as opposed to saving them as AutoCAD drawing files (.dwg) or exporting them as Arc View shapefiles (.shp) or Microstation design files (.dgn). The reason for this is that .dxf is the only format that maintains the 3D spatial relationship created by connecting two points at varying elevations when exporting them from AutoCAD MAP R3 into Arc View 3D Analyst.

Fifty-nine control points were set up around and within the site from where each feature was recorded (fig.5). These points were established by resecting off of two known points beginning with the origin point and the point marking the contact area where the north line touches the Hanging Remnant. Side shots were collected every 10 cm vertically while the horizontal mapping resolution ranged from between 25-50 cm depending on the degree of topographical variation of the feature. For example, a wall of breccia with small inclusions is relatively unifonn so fewer points were taken to represent it, while a dolomite roof block is characterized by undulations and crevices and was respectively assigned more points due to its complexity (fig.6). These mapping intervals were chosen to be consistent with Brain's excavation levels. Aline level was used as a rough guide for mapping. For areas that were inaccessible, the vertical and horizontal measurements of the theodolite were monitored and points remained unrecorded until these readings were consistent with the appropriate intervals. This degree of accuracy was sufficient enough to allow a detailed image of the terrain to be produced from the interpolation of the survey points. Attributes for the points were recorded, including feature descriptions, geological material, sediment color, degree of calcification, cave area, Member (strata), and the cardinal direction that the feature is facing. This last attribute was essential for the interpolation of the data using Voxel Analyst, our principal software, for reasons discussed below.

Interpolation of 3D data and the limitations convential GIS data models

of

A range of GIS and related software packages, including Surfer 7, Arc View 3.2, ldrisi 3.2, Slicer-Dicer and GRASS 4.0 were tested to see how their interpolation algorithms would handle point data that had multiple z values for the same x,y coordinate pair, representing the natural geological overhangs at Swartkrans (fig.7). After many attempts, it was realized that none of the interpolation methods were suitable for processing this type of data. Due to the data structures of these software packages, the resulting interpolated image contained severe errors making the 3D reconstruction of Swartkrans highly inaccurate in certain areas. Some of these software packages, such as Surfer 7, prompted the user to choose a calculation parameter, such as minimum, maximum, or average, for points possessing multiple z values. Other software packages, such as Arc View 3 .2, failed to grant the user the option to select the points containing du;licate x, y coordinate values that were to be included in the interpolation. In both cases the resultant output was characterized by data spikes in places where overhangs are present at the site, such as below the Hanging Remnant (fig.8). This is not surprising because the data model of a conventional GIS is structured in two dimensions where features are assigned x and y spatial coordinates with an attribute defining some aspect of the feature, such as soil type. Two-dimensional mapping results in planar surfaces where each data layer represents some characteristic found on a single surface. When the attribute signifies an elevation or a measured value of some sort then the data model is considered to be surficial, or representing 2.5D data. Surficial data can be visualized in three dimensions although it lacks 3D topology. A classic example of this is a Digital Elevation Model where height is considered an attribute (Raper 1989; Turner 1989). Unfortunately, up to this point archaeological applications have been limited by the constraints

The survey data were stored on a memory card that was transferred daily from the total station to a Wild GIF 10 data collector. The data were then downloaded from the data collector to a Dell Inspiron 3200 laptop using a data transfer program called Convtran. The point IDs, X- coordinates, Y-coordinates, and Z-coordinates were saved in text file format. Finally, these space-delimited files were imported into MS Excel where field headings were assigned and attributes were entered manually. The application of photogrammetry was not considered for this project due to the lack of proper equipment and training and to the site being characterized by small recesses that are easier to

115

of these two types of data models. Yet, excavated material exists on multiple artificial surfaces overlaying one another in 3D space, making these data models insufficient. As a result, spatial modeling software was used instead of traditional GIS algorithms to produce an interpolated image containing all overhanging features. For an extensive overview and discussion concerning the progressive stages and limitations of representing multidimensional data in archaeology, refer to Harris and Lock (1996).

Shephard's Method algorithm, an inverse-weighted approximation. This was the most appropriate interpolator since it is ideal for rough and unevenly distributed data and it satisfies the Maximum Principle in which interpolated output values will not exceed the range of the original input values (Intergraph 1994). The output represented an accurate reconstruction of Swartkrans, including all existing overhangs. The final model was exported from Voxel Analyst as a .dgn file, opened using the CAD Reader extension in Arc View 3 .2, and finally converted to a .shp file.

Constructing the 3D Model of the Outer Cave

A mistake realized in hindsight, after the survey was completed, is that no survey points were recorded on the ground level of the site since there isn't much variation in elevation. This resulted in the appearance of a "bulbous floor" in the final reconstruction of the site due to the scarcity of points representing these areas (fig.10). The ground level, fortunately, represents areas that may be excavated and recorded at a later date and so the data can always be incorporated into the GIS in the future.

Voxel Analyst, developed by the Intergraph Corporation, provided the solution to the problem mentioned above since it recognizes three independent axes (x,y,z). The software played a crucial role in this project and was responsible for making the 3D reconstruction of Swartkrans possible. Voxel Analyst is a data visualization and analysis tool that is designed to characterize subsurface data and to enhance the understanding of multi-dimensional data relationships within a 3D volumetric data set. A voxel, or volumetric element, is a three-dimensional version of a pixel, depicted as a cube. The voxel data model was introduced in commercial software that was developed for the purpose of oil and gas exploration It is currently used for a wide variety of applications, from reservoir engineering to mining exploration (Harris/Lock 1996; Turner 1989). Utilizing this type of software resulted in an innovative way to accurately interpolate a terrain model of an archaeological site in true 3D space.

Converting the excavation notes into digital format Brain's excavations lasted from 1979-1986 and produced several volumes of excavation notes outlining in meticulous detail the geological contents, and, in some instances, the fossil content of each excavated grid square. Converting information contained in these notebooks into digital fom1 was essential to rebuild the site as it existed prior to Brain's excavations. By doing this, the user of this system will be able to view specimens within their associated strata once again and to conduct a virtual re-excavation of the site.

In order to model the existing terrain at Swartkrans using Voxel Analyst, the site had to be viewed as a solid block. Thus, the solid earth as well as the void areas representing air that once contained the excavated material had to be assigned unique numeric values respectively. This allowed the software to acknowledge both features as two distinct phenomena interlocked in three-dimensional space (fig.9a). Since a reflector card was not used to collect survey points in mid-air with the laser during the initial survey, the air component was assigned using another method. Essentially all of the survey points representing the cave surface were duplicated in MS Excel and repositioned 1 cm (the distance between the center of two adjacent voxels) away from the surface in the appropriate cardinal direction in order to create the outer boundary of the air mass (fig.9b ). This made it possible for the software to identify the interface between the cave structure and the air and to create a thin surface located where each voxel adjoined each other (fig.9c ). This procedure eliminates the problem of spatial inexactitude that is usually inherent in voxel-based systems due to the voxel size and the lack of boundaries between spatial entities (Jones 1989).

The excavation diagrams were digitized using a Microscribe Digitizer, though any digitizing would have sufficed (fig.I la). Initial calibration was performed using Inscribe software to define the origin, coordinate axes and scale. Most of the diagrams were drawn on a standard form where each side of the grid-square represented 80 mm (fig. llb ). All of the diagrams were rescaled to convert their original dimensions to equal lm 2 • Sketched drawings were also included to provide infonnation where standard diagrams were unavailable (fig.11c). Although these drawings are less accurate representations, it is better to have some information present in the GIS for these areas than not to have any at all. Each diagram was digitized, reprojected to its appropriate grid square in true geographic space, and stored on a unique elevation layer inAutoCAD MAP R3. The elevation layer of the diagram was assigned based on the bottom-most level of the 10-cm spit from which it was drawn. For instance, if a grid square was excavated at 80 - 90 cm below the datum, then the data in that area were assigned an elevation of -0.900 m due to the fact that each diagram was drawn after the material was removed from the grid square. Each elevation layer with associated excavation data was exported as an ArcINFO coverage and brought into PC ArcINFO to create topology. The attributes were then extracted from the excavation notes and entered in MS Excel to compile a table 1 of geological features. The digitized diagrams were finally imported into Arc View 3.2 where the attributes were linked to the appropriate feature using unique ID values selected as the join field or

Man-made features at the site, such as steps, were initially mapped and then eliminated from the survey data prior to interpolation of the natural surface. The text file representing all the survey points was brought into a text editor. A carriage return, necessary for the software to recognize the file, was placed after the last line of the data. The file was saved in .smp format and imported into Voxel Analyst. The 14,489 (14,885 minus the man-made features) points were interpolated using the

116

The amalgamation Analyst

primary key (fig.12). These polygon themes were then clipped based on the appropriate elevation layers extracted from the 3D model. This procedure combined the new data with the old and ensured that the existing geological infmmation of the site was taken into account before the finds were inputted into the system. At times, entities derived from the excavation note descriptions, as opposed to the diagrams or sketches, were deleted as a result of the clipping process. These features were reconstructed later using Arc View editing functions.

Once these three sources of Swartkrans data were either created or converted into a usable digital format they were displayed together using the 3D Analyst extension for Arc View. All of the data was spatially sound and overlayed each other perfectly (fig.13). Now that these themes were brought into 3DAnalyst, the three-dimensional capabilities of the software needed to be assessed.

Creation of the fossil and artifact database

Although 3D Analyst would be expected to have threedimensional analytical capabilities based on the name, the data model of the software remains surficial. The Swartkrans data can be displayed very nicely within the module, but proper 3D analyses and functionality, such as nearest neighbor and buffering operations, cannot be conducted due to the lack of 3D topology. For example, when the user wanted to know how many fossils lie within 10 cm of a dolomite block, 3D Analyst would only output the points that are within a 10 cm area of dolomite on the same horizontal plane not taking into account the points that lie above, below, or at angles from the block. This is a serious problem since it limits the integration of these datasets to visual analysis, besides the conventional 2D and 2.5D operations.

Throughout the 1999 project year, one of the authors (DJD) compiled a database of the fossils and artifacts found at Swartkrans. This task was carried out at the Transvaal Museum in Pretoria, South Africa, where all relevant infonnation previously gathered was entered on computer using MS Access. The finds entered were only those from Members 1, 2, and 3 that were assigned accession numbers beginning with SKX. The resulting database contains almost 22,000 records with attributes recorded for numerous fields 2 • The fields created for the survey and excavation data were kept consistent with those listed above, when applicable.

Point representation three-dimensions

of the data within ArcView 3D

of the fossils and artifacts in 3D buffering operation and potential applications

Although the locations of many of the finds were recorded only to the grid system and spit (level), it seemed more beneficial to represent them as individual points within the GIS. A point theme allows the user to visualize the fossils and artifacts of interest in and about the general area of their discovery and to identify clusters of certain fossils more readily without having to deal with the confusion of having each grid square representing multiple points at a time.

In an attempt to bypass this obstacle, an Avenue script was created that is successful in creating three-dimensional buffer zones. The script, however, is limited to the buffering of point features. An Avenue script (sphere3D_creator) 5 that can create a spherical graphic around a user-specified point, at a defined radius, was downloaded from the ESRI web page 6 • The script was combined with another developed independently that uses a 3D Euclidean distance calculation in order to measure the distance from a user-defined point to all other points in the dataset. The calculation is as follows: SQRT((((xcoordinate of specified point - x coordinates of all points)2) + ((ycoordinate of specified point-y coordinates of all points) 2 ) + ((zcoordinate of specified point - z coordinates of all points) 2 )))

The file heading informs the browser the file language: xml, version 1.0. The encoding used is the one defined by utf8 abbreviation. the transformation fom1at lJCS (Universal Multiple-

166









http://www.we.b3D.org/

TaskGroups/x3d/translation/README.X3D-Edit.htrnl" name="generator"/>

(sk_ipp_inq

code)

The file header includes meta tags, with various additional informations about the file, the author, the date of creation, etc.



11

There is a definition of the viewpoint, then the end of the scene and of the X3D file.



.11

' 56.0

50.0 40.0

5·1 .0

30.0

52.0

20.0 10.0

,o.o

0.0

1650

Figure 9. Example variable plot, Nlh (males only, right bank), predicting bank model. Note that the trend is generally increasing, although with the large uncertainties (+/-1 standard deviation), such observations are not conclusive. The x-axis is the midpoint date as presented in Table 5 and the y axis is the mean cranial measurement for that time period. The line is the best fit regression through the average.

Figure 7. Summary of results, archaeological database, 0.75 threshold. Labels as in Figure 6.

...- ····-·-··-·-····-·-·-·--·----·-·

80.0 70.0

f

60.0

500

l

--··-·-

f

f

40.0

•

-·---

f

f

l

f

l 766

1S l7

Figure 9. Example variable plot, Nlh (males only, right bank), predicting bank model. Note that the trend is generally increasing, although with the large uncertainties (+/-1 standard deviation), such observations are not conclusive. The x-axis is the midpoint date as presented in Table 5 and the y axis is the mean cranial measurement for that time period. The line is the best fit regression through the average.

Figure 7. Summary of results, archaeological database, 0.75 threshold. Labels as in Figure 6.

90.0

l 70 4

4i'.O

30.0

45,0

20.0

43.0

10.0

0.0

.j>,

~'\

SI! ~,,, .,,. ~

Figure 8. Summary of results, archaeological database, 0.60 threshold. Labels as in Figure 6.

,It' '